act essay length

The ACT Essay - An Introduction

So, the big question is this: should you even care about your ACT essay? If it’s optional, and it doesn’t really affect your score, then why would any sane person take this thing at all?

Three answers to this question:

• It depends on the colleges . Many competitive universities recommend the “ACT with Essay” when submitting standardized test scores. Because most of you will be applying to the top 100 universities in the United States, I expect every student taking an AP Guru course to take the ACT Essay. It’s up to you to research your colleges of choice and figure out what their policy on this will be.
• It can be used as an indicator to “verify” your actual application essay. Unfortunately, there’s a growing trend of students getting “help” on their essays. Getting guidance and editing is fine, but a lot of essays now are basically written by a professional and submitted as student work. Many colleges will use this essay to see if your writing style is consistent. If you write like Hemingway in your application essay and like a 3-year-old child on your ACT essay, they’ll know something is up.
• It’s so easy to ace that you might as well take it. The ACT Essay is easy to prep for. Although the essay may seem intimidating right now, after you learn our AP Guru essay strategies, you will find that the ACT Essay is perhaps your favorite and easiest part of the ACT.

So honestly, why not just do it? It takes 40 minutes to write - that’s time you’d probably spend watching Netflix on the Saturday of your test. And because it’s at the end of the test, it’s not going to eat into your mental energy.

Overall, my answer would be: just learn to write this essay well because it’s easy, it can’t really do any harm, and it COULD do some good.

Getting a good ACT essay score is so easy that there is no reason for you to avoid it. The ACT essay has nothing to do with your level of writing talent, and everything to do with strategy and knowledge of the grading system. I don’t care if you’re “a math person, not a writer.” It doesn’t matter. In fact, some of the best ACT essay writers I’ve ever worked with were great at math and terrible at English.

So, now since you have decided to give the essay, it’s important to understand what it actually is about.

Basics of the ACT Essay

The ACT will present an issue with context and accompany it with 3 perspectives . The prompt will ask you to

• “Evaluate and Analyze” the three perspectives presented
• “State and Develop” your own perspective
• “Explain the Relationship” between your perspective and the given perspectives

You need to do this in the four lined pages provided to you.

How is the Essay Graded

Once you have sent the essay in, two readers each will give it a score from 1-6 in the following categories:.

• ideas and analysis
• development and support
• organization
• language use
• Ideas and Analysis You’ll need to make sure you address and thoughtfully discuss multiple perspectives on the issue. Furthermore, a top-scoring essay should make sure to address an issue in its full complexity. Rather than trying to simplify the issue, make sweeping generalizations, or rely on assumptions, a top scoring essay acknowledges the strengths and weaknesses of the proposed position.
• Development and Support Make sure that you support your point with lots of compelling evidence, evidence that fits well together into a unified whole and includes things like exceptions, edge cases, and limits to show that you know specifically where that evidence or logic does and doesn’t apply.
• Organization This is an easy one. Harnessing the writing skills you’ve learned throughout school, you just need to lay out your overall essay in a logical way. By following our template, you should be able to receive top scores in this category with ease.
• Language Use Use rich, interesting vocabulary to show your mastery of the English language. Use sentences that have different lengths and structure, with a professional tone, and make sure to proofread so that there are few to no errors present.

In the chapters to follow, we’ll dissect the above into concrete “checkboxes” that you can cross off one-by-one as you write your essay. Now, you need to understand how the essay graders assign points for these different categories.

The Grading Procedure

For every ACT session, the organization administering the ACT receives hundreds of thousands of essays from students across the world. How does it manage to grade every single one of them in 2-3 weeks?

First, it hires a bunch of teachers who presumably know something about grading papers. But that still does not solve the problem. Next, it has to standardize the grading process so that scores remain consistent across the board and the teachers can get through the sheer number of essays quickly.

The organization standardizes the process by asking graders to put their own opinions aside and solely correct using set guidelines. Any flashes of your literary brilliance will be glossed over in a ruthless grading procedure that only cares about whether your work matches the standard model.

Your job, then, is not to write a masterpiece. This isn’t your IB English class. The teacher won’t be looking at every word and dissecting all your punctuation marks. Instead, your job is simply to write something that fits the guidelines. That’s it. It doesn’t have to be innovative and it doesn’t have to be in your voice and style. Just give them what they want.

Length of your Essay

One question that all my students ask me: Does the length of my ACT essay affect my score? The short answer is: YES. Longer essays usually get better grades.

If you aren’t filling all four pages of your booklet, it’s not the end of the world. But if you’re barely filling up two pages, that’s going to be a problem.

However, there’s a risk here: the longer your essay is, the more chances there are for you to go off-topic, break up your structure, or put in random sentences that kill the flow of your essay. You especially don’t want to go off topic that’s the ultimate essay score destroyer. By planning your essay in advance and sticking to the template you’re about to learn, you will ensure you do not go off-topic.

Step by step process

How can you crank out an essay in 40 minutes here’s my step by step process when i first open the writing section screen:.

• Read the informational paragraph at the beginning of the prompt. It only takes 30 seconds and it gives you context for what you’ll be writing. It might also help you think of the examples you’ll use in your essay.
• Read the three perspectives.
• Choose one of the three perspectives to be your own stance.
• Come up with an example for the perspective you support and one each against the perspectives you disprove of.
• Don’t read the essay task - it’s always the same. You should know from practice what the assignment is.
• At this point, it’s time to start your essay . You should start typing the essay within 5 minutes of the starting time.
• It’s pretty smooth sailing once the essay starts coming together in your mind. Just follow the template for the body paragraphs and the conclusion.
• Always have a conclusion . The graders will punish you if you don’t have one. I’d rather cut one of my body paragraphs short than run out of time before the conclusion.

The topics you will write about will be different, but the overall structure and the way you write will be identical every time.

Flaws to Avoid

As I keep repeating to my AP Guru students, its often more important to know what not to do then what to do . So, the following are list of points that are ACT Writing point killers and should be avoided at all cost.

• Not writing about the topic presented . If you write about a subject other than the one presented by the ACT, you’ll automatically get a 0. Write your essay about the prompt provided, please
• Not developing a clear thesis. Selecting a point of view and then supporting it with evidence is the entire point of writing this essay - if you don’t have a clear thesis, you’re never going to get a good score.
• Lack of a coherent “thread” throughout your essay. Put simply, if your essay is all over the place, you’re going to get a terrible score. If you randomly skip to certain points, or omit evidence, or keep switching from A to B to D to A again, you’re in trouble. You need an intro, three body paragraphs, and a conclusion, and you need them in that order.
• Avoid Circular Reasoning. “Circular reasoning” is any form of argument in which you try to prove a point by simply repeating it with different wording. Your using different words, but I’m still saying the exact same thing.
• Providing irrelevant examples or evidence . Irrelevance will kill your ACT essay score, and this is where most “good writers” get in huge trouble - they try to be creative, but they end up being incoherent. Your evidence and examples can stink, but they need to be related to the point you’re trying to prove.
• Bad grammar. Remember: the essay is grouped in with the English section. If you make tons of grammatical errors, you’re going to lose points. A single mistake here or there won’t affect your score, but once you hit a critical mass of mistakes, you’re cooked.
• Lack of transition between paragraphs . Each paragraph you write will be dedicated to a certain idea. There’s nothing worse than toggling between ideas with no transition whatsoever.
• Errors in spelling and vocabulary. If you constantly misspell words or use them incorrectly, you’ll lose points. Once again - a couple tiny mistakes won’t matter, but there’s a critical mass you can’t afford to hit.
• Illegible handwriting. If I can’t read it, I can’t understand it. If I can’t understand it, it makes no sense.

You Might Also Like

Co-ordinate Geometry

COORDINATE GEOMETRY is the study of geometry on a plane using a coordinate system. This plane is called a COORDINATE PLANE, which has scales of measurement along the x and y-axes.

Transitions

Transitions are one of the most common and trickiest topics tested on the SAT Writing + Langauge and Language Test.

Hard ACT Questions

In every section on the ACT, there are some questions that are more difficult than the others. Science is no different. Each of the ACT Science passage tends to have 1-2 hard questions that can appear at any point in the passage.

Free Resources

Everything You Need to Know About ACT Writing

The ACT Writing test, or essay section, tests your ability to read different perspectives on a topic and present your own opinion in writing. It’s the optional last section of the exam, meaning that you don’t have to take it if you don’t want to–although there are many colleges and universities that require or recommend it for admissions. Here’s what you need to know about ACT Writing in a nutshell!

What is the ACT Writing section?

Here are the need-to-know facts about the ACT Writing section :

• It’s a 40-minute test to complete with pencil and paper
• You’ll get one essay prompt about a debatable topic and three different perspectives on it
• evaluate the three different perspectives
• present your own perspective (which may agree in part or in full with any of the provided viewpoints)
• explain the relationship between your viewpoint and the provided ones.

Do I have to take ACT Writing?

Nope! It’s not a required part of the exam. You can think of it as an add-on that, yes, also costs a little bit more. But whether the test requires the essay and whether colleges do are two separate issues. While you can submit the ACT without writing to many colleges and universities, others (including most super-competitive colleges) will ask you for it. Because the list of which colleges require ACT writing is constantly changing, it’s best to check with the individual schools you’re thinking of applying to before you register.

With that said, some students do take the exam before they know where they’ll be applying to college. So in this case, is it better to do the ACT test with or without writing? If you’re really uncertain, take it with writing. This will ensure that you don’t have to go back to the test center because you’ve decided to apply to a college that does require it in a few months! Yes, it requires some extra prep–but it’s worth it to keep your options open.

How does ACT Writing scoring work?

If you’ve been prepping for the ACT overall, you’re probably very (overly?) familiar with the 1-36 scoring scale by now. But this isn’t how the ACT essay is scored! Because it’s not a multiple-choice section, the essay has its own unique scoring. Take a look!

How ACT Writing Is Scored

The writing portion of the ACT is scored by two graders, each of whom will assign you a writing score of 1 to 6 on four different writing “domains.” These domains include:

• Ideas and Analysis: does your thesis statement and writing examine the implications of the issue with clarity and nuance?
• Development and Support: do you use well-developed and specific examples to support your thesis?
• Organization: does your essay have a strong structure and a logical flow?
• Language Use and Conventions: do you use correct grammar, word usage, mechanics, punctuation, and sentence structures?

Your total points added up between these two graders are converted from these domain scores to a scaled score of 2 to 12, which is the final score you will see on your score report.

Getting Your ACT Writing Scores

So you’ve done it! You sat down, you wrote your essay…now what? When will ACT Writing scores be available? Well, you’ll get your multiple-choice scores in around two weeks, and you’ll get your ACT Writing scores (and full score report) about two weeks after that. So anticipate waiting around a month to get the total view of how you did on the ACT essay.

Can ACT Writing hurt your score?

No! Not, that is, if you’re talking about the composite ACT score on the 1-36 scale. This is an average of your 1-36 scores in each of the four multiple-choice sections (Reading, Math, English, and Science). It doesn’t take your essay score into account. In other words–does the ACT writing score affect composite scores? It doesn’t.

With that said, the essay score will affect your ELA subscore on the exam. (You can see what this looks like on a score report here .) But colleges and universities are far more interested in your sectional and composite scores.

What essay score is good?

Generally, an 8+ is an excellent ACT Writing score , but it depends on where you’re applying. Ivy-League-bound students should aim for 9 or above !

Can ACT Writing be superscored?

Nope! The ACT superscore is a way of combining your highest multiple-choice scores from different sections and different test dates to give you the highest possible combination. But it only affects your composite score . Because the essay isn’t part of your composite score, it’s not part of the superscore, either.

What happens if I mess up on the writing section?

If you’re wondering if you can cancel the section after they’ve taken it, unfortunately, the answer is no. Once you’ve written the essay, your score will appear on your score report. But if you change your mind about taking the writing section of the ACT, you can cancel it before you get into the testing room.

If you feel that you didn’t do your best, or you get a lower score than you’d like, can you retake ACT Writing? Definitely! But for the moment, ACT sectional retesting is on hold—and it may or may not include the ACT writing test being taken separately. To retake ACT Writing, in other words, you’ll need to retake the entire exam…for now.

How to Improve Your Writing Score

So how do you boost your ACT essay score? Take a look at these tips!

✅ Practice planning and writing essays on practice ACT essay prompts. Although writing full essays (preferably as part of ACT practice tests ) is the best practice, ten-minute outlining sessions in which you plan out your essay (like you will do on the test) can go a long way in helping you learn how to quickly generate and organize your ideas.

✅ If grammar is not your strong suit, check out our post on ACT grammar rules . Not only will this help you improve your grammar for the ACT essay, but it also will help you with the ACT English section !

✅ Share your writing with the strong writers you know and get feedback from them. Have them score your practice essays using the ACT rubric .

✅ Review the sample essays on actstudent.org so that you can get a sense of what kinds of essays get which scores. This can be incredibly helpful!

✅ Learn about current events and form your own opinions on them. Engage in lively debates with your friends and family so that you can practice supporting your opinions and anticipating opposing arguments!

Want even more? Read Your Magical Guide to Scoring a Perfect 12 on the ACT Essay , which will guide you through all the basics (if you’re a beginner) and how to improve your score (if you’ve already taken the test). There, you will find lots more writing tips and strategies that will help you show the ACT Writing test who’s boss. Happy Studying!

Rachel is a Magoosh Content Creator. She writes and updates content on our High School and GRE Blogs to ensure students are equipped with the best information during their test prep journey. As a test-prep instructor for more than five years in there different countries, Rachel has helped students around the world prepare for various standardized tests, including the SAT, ACT, TOEFL, GRE, and GMAT, and she is one of the authors of our Magoosh ACT Prep Book . Rachel has a Bachelor of Arts in Comparative Literature from Brown University, an MA in Cinematography from the Université de Paris VII, and a Ph.D. in Film Studies from University College London. For over a decade, Rachel has honed her craft as a fiction and memoir writer and public speaker. Her novel, THE BALLERINAS , is forthcoming in December 2021 from St. Martin’s Press , while her memoir, GRADUATES IN WONDERLAND , co-written with Jessica Pan, was published in 2014 by Penguin Random House. Her work has appeared in over a dozen online and print publications, including Vanity Fair Hollywood. When she isn’t strategically stringing words together at Magoosh, you can find Rachel riding horses or with her nose in a book. Join her on Twitter , Instagram , or Facebook !

View all posts

More from Magoosh

Leave a Reply Cancel reply

Your email address will not be published. Required fields are marked *

How to Write an ACT Exam?

Your high school GPA may play an important role in your college admission, but the ACT exam may have the final say. What is an ACT exam? The ACT is an entrance exam that most colleges utilize to decide on a candidate's admission. The main purpose of the writing test is to disclose and measure the applicants' readiness for college studies. 

The admission officers review the test score along with your high school GPA to make their final decision. Of course, the classes that you took, extracurricular activities, letters of recommendation from your teachers, and your application, where you discuss the high school experience from your own perspective, will be reviewed as well. While the importance of the ACT scores may vary from college to college, the higher the score, the more options for the college enrollment you have. 

The ACT exam consists of four sections that include English, Reading, Math, and Science. Aside from that, the prospective candidates are given forty minutes to do the writing test. Writing an essay can be quite difficult; that's why you need to prepare for it. The good news is that ACT essay examples are available online, and you can view sample in this article as well. So, basically, that's what the ACT exam is about. But let's dive into the details with our dissertation help services team. 

What is the ACT Exam?

The ACT exam 2022 consists of several sections that reveal your knowledge of various school subjects. The ACT exam stands for “American College Test”. It's a multiple-choice test where you have a number of questions and several options of answers. And you need to be ready to give it almost four hours. 

How long is the ACT exam exactly? You have two hours and fifty-five minutes for the multiple-choice test and forty minutes for the writing section. The ACT exam 2022 length gives you enough time to choose the correct answers and dive into your essay, which is quite complex. Basically, in the writing test, you will have to expand on the provided important topic and three different perspectives on it. 

When writing a coherent essay, make sure that you show your abilities in several core aspects. You need to be capable of generating ideas and analyzing them from various perspectives. It's also crucial that you provide pieces of evidence that support your viewpoint. The essay must be organized logically, so the reader can easily slide through the narrative. Finally, proper written language use and conventions are also extremely important. 

The Structure of the ACT Writing Test

If you are planning to pass the ACT exam 2022, you need to know the structure of the exam. First, let's look at the time you're given for completing each section, as well as the number of questions each ACT section has. What is the ACT exam, if not an exercise on time management? You have to plan basically every minute while you're taking it. Let's divide the ACT exam length for every section:

How to Write an Act Exam

All in all, you have 3 hours and 35 minutes for the ACT exam, meaning that you have just enough time to go through the test once and mark all of the right answers. Aside from it, students have a ten-minute break after the Math section and five minutes rest before they start writing the essay. Try doing an ACT practice exam and actually see when you might need breaks. You've already learned the most important aspects of the written ACT exam scoring. Now, let's take a look at each section of the multiple-choice test. 

English Test

The ACT English test consists of 75 questions, and the 45 minutes given for it means that you'll have between 30 and 40 seconds for each writing. This section can be divided into three parts in terms of content:

Conventions of Standard English 

This is by far the largest part of the test, as more than 50% of the test is devoted to the conventions. The questions focus on grammar, syntax, punctuation, and the correct use of words. Here you may be asked to choose the correct verb or tense. This part focuses on the general rules of English, so you don't have to write essays, sharing your own perspective on world problems or books.

Production of Writing

Around 30% of the test is devoted to the production of writing. Here, students are given passages from the various texts to check how good they are at seeing the big picture. You must choose the correct answer based on the tone or purpose of the passage. Questions in this section also concern the literary style and strategies used by the author. 

You may need to also know what is diversity essay while preparing for the exams.

Knowledge of Language

This is the smallest part of the English test, as its share ranges between 10% and 20%. This segment focuses on style, tone, and precision. 

The Math test may seem to be a bit too easy at the beginning. But that's because it's arranged in order of difficulty. Basically, you can say that out of sixty questions; the first twenty are quite easy. The next twenty questions are of medium difficulty, while the last twenty questions are the most difficult ones. Check an ACT practice exam to see at what point you really start to struggle with the questions so that you have an idea of what to review. Now, let's see the percentages that each math topic has in the test:

• Pre-algebra - accounts for 20%-25% of questions;
• Elementary algebra - ranges between 15% to 20% of questions;
• Intermediate algebra - up to 20% of questions;
• Coordinate geometry - takes 15%-20% of questions;
• Plane geometry - around 25% of questions;
• Trigonometry - ranges between 5% to 10% of questions;

The good thing is that the ACT exam allows using a calculator, so you don't have to waste a lot of time if you are not the brightest in terms of calculation. 

Reading Test

The reading test has a lot in common with the production of writing questions in the English test. In this section, students are given passages from texts and questions about the passages. You may be asked to identify the author's attitude toward the main subject of the passage. The meaning of this or that word in the context of the passage. The main idea of the body paragraphs. The meaning of the phrase or the function this or that sentence serves within the text. The important part is that you don't have to develop your unique perspective of the text - this is the part of the test that measures your ability to analyze given information without adding your own perspective on the issue.

Science Test

The Science section includes topics like biology, chemistry, space and earth sciences, and physics. You may be asked to pick the right conclusion from the options you have for graphs and diagrams. The section includes questions on conflicting viewpoints. Here, students are provided with a description of a certain phenomenon with several hypotheses on its explanation. Then you need to choose which assumption reflects the notion provided in the questions. 

How to Improve Your Writing Skills?

Considering the ACT exam dates 2022, you still have time to practice the ability to discuss ideas and convey arguments with clarity - improve your writing skills for the essay. You don't know the topic you’ll get, so deciding to buy an essay paper won't be that helpful. But there are a few things that you can do to improve your writing for the ACT exam. 

• Read and write - if you do it frequently, you won't have problems with the test;
• Get familiar with current social issues and various opinions on them;
• Taking debate classes will prepare you for the essay;
• Try writing extracurricular activities, like creative writing clubs;
• Share your writing with others;

Basically, improving your writing can be shaped into a simple formula: read-write-repeat. Read how other people write and try writing yourself. That will master the style and improve your ACT exam scoring. You can try blogging to master your skills. As for the organization of your writing, you can always seek advice from your English teachers and generate productive ideas. Asking your classmates and friends what they think of your writing can also help greatly. Just remember that no one is born with perfect writing skills. You can achieve it only through practice. 

Use our free Plagiarism Checker!

How to Write an ACT Essay?

After all, despite the complexity of the ACT exam 2022 multiple-choice tests, the essay task remains the most difficult part of it. You might have checked ACT essay examples, but writing one yourself in a short time of 40 minutes can be quite stressful. You need to have a strategy for writing. So, here are the steps that can help you. 

• Devote around ten minutes to planning your essay. 
• Think of the way you're going to organize it. 
• Read the one writing prompt and the multiple perspectives, and choose the one you're going to work on. If you check an example ACT essay, you know what to do. 
• Brainstorm the ideas and analysis, as well as supporting pieces of evidence you can provide. 
• Picture the structure of your essay, and then devote 30 minutes to writing it so that you will have just enough time on the test day. 

You want to know how to write Common App Essay Prompts 2022-2023? We encourage you to check out the fresh topics in our article.

ACT Essay Sample

Finally, we are getting to something you can build your essay on. So, you are to get a prompt with three different perspectives on it, and you need to choose which one of them you're going to write your essay, just like the example ACT essay. Let's imagine that the writing prompt you get will be on cancel culture, as shown in many of last year's ACT essay examples.

"Cancel or callout culture is a tradition of removing the support for certain individuals or their work based on their actions or opinions that can be considered objectionable. Traditionally, individuals are called out on social media, which leads to the general awareness of their perceived offense. In most cases, canceling takes the form of pressuring organizations to prevent the subject of the canceling from public appearances. When it comes to business, it can include boycotting their products" — that's what the writing prompt may look like. 

So Much Homework That You Don't Have Time to Prepare for the ACT?

Delegate the part of your homework to our paper writers for hire , and spend your time preparing for the ACT exam. Let the professionals do the paperwork while focusing on important things!

Now, let's check the three perspectives that you can be provided to defeat the ACT exam scoring system:

Perspective 1:

Cancel culture allows marginalized people to hold their tormentors accountable via public opinion when the justice system fails them. Movements like #metoo allowed numerous women and men to call out their abusers. The cancel culture leads to a safer world, free from abuse of various forms. 

Perspective 2:

While the cancel culture allows people to call out their abusers, we cannot disregard the fact that certain people can use it for their own benefit. The culture leads to people being canceled based on proofless claims, which turns the anti-abuse campaign into witch-hunting. 

Perspective 3:

Cancel culture may lead to a change in power but is ultimately unproductive when it comes to social change. As soon as it became big, it turned into simply casting stones at each other. While we may think that it leads to social justice, it's just engaging in the social media frenzy that doesn't affect real life. 

So, here you have the complex issue and three perspectives on it. First, you need to analyze and organize ideas of everything you know about the cancel culture. Think about the examples that you are acquainted with well. Then, examine each of the perspectives presented to you. Think which one is closest to your own thoughts about the issue. Alternatively, you can choose the perspective that you can support with more pieces of evidence. 

You start your essay with an analysis of the cancel culture as a whole. Then you should move to one of the perspectives given. There, you can expand and convey arguments on the topic by providing your own viewpoint. Your viewpoint must be supported by pieces of evidence. Regardless of the perspective you've chosen, in the case of cancel culture, the Harvey Weinstein and Depp vs. Heard cases will surely have something to do with it. Then you finish your essay with a logical conclusion. 

Don't forget that there must be a logical transition between the passages of your essay. The next passage must expand on the information provided in the preceding one. Your viewpoint must be followed by the supporting evidence. Here’s a sample of an essay that uses perspective two as the foundation: Here you go the ACT exam sample:

Is the ACT Essay Optional?

If you checked the ACT exam dates 2022, you have several months to prepare yourself. The first exam will occur on September 10, 2022, while the last one in the calendar year will be held on December 10, 2022. The biggest question that bugs students is whether the essay is optional. Yes, it is.  Moreover, not every college requires applicants to write an optional ACT essay.

Another important thing is that the total ACT exam scoring is not affected by your essay score. Some colleges think writing the ACT test gives candidates a more rounded score. So, all in all, the essay score matters only together with your multiple-choice results. So, why bother writing the essay at all? 

If you are highly competitive, you should try your hand in the ACT writing section. To prepare for it, you can follow the instructions given in this article and read through the ACT essay examples. If you are a mid-range student, you can opt for the writing test, but you should stress out about it, considering that many colleges do not require completing this section. Less competitive students should avoid the ACT writing test and stick to the multiple-choice test. The only reason they should take it is if the college board requires the essay for admission. 

How to Pass ACT with Flying Colors?

Now that we have discussed the ACT exam meaning, you can realize that it is not an easy task, but it's not that difficult either. Given all the information provided here, you know what to expect from the multiple-choice test and the writing portion, and you already know the ACT exam scoring system. ACT exam dates 2022 are approaching, but there is nothing to be scared of. Check the ACT essay examples , read through test tips once again, solve a few training tests, think about the logical structure of your essay, and keep the scoring rubric in mind- now you are ready to go and take the test without fearing the final score!

Let the experts do the work while preparing for the ACT!

You just can't get it all, so don't compromise and get the best of your last few months of school! Graduate with a great GPA by delegating some of the tasks to our experts while you'll be spending time getting ready for the big test! If you need essay help , just give us your requirements and we'll give you a hand asap.

Perhaps you need some nursing or psychology essay writing service ? Do not hesitate to contact us! Our team is always ready to help.

Pass Your Exams with Flying Colors!

Get help from our experienced team of the writers

Related Articles

• Plan for College and Career
• Take the ACT
• School and District Assessment
• Career-Ready Solutions
• Students & Parents
• Open Search Form
• The ACT Test

• Registration
• Test Center Locator
• High School Codes Lookup
• Photo Submission Requirements
• Standby Testing
• Accommodations and Supports
• Free ACT Test Prep
• Official ACT Subject Guides
• The Official ACT Prep Guide
• ACT Kaplan Test Prep Suite
• Rescheduled Test Centers
• CAS Calculator FAQ
• College Codes Lookup
• How to Send Scores
• How Schools Use Scores
• ACT vs. SAT
• When to Take the ACT
• Understanding Your Scores
• Your Test Questions and Answers

Other ACT Services and Products

Writing Sample Essays

Write a unified, coherent essay about the increasing presence of intelligent machines. In your essay, be sure to: 

• clearly state your own perspective on the issue and analyze the relationship between your perspective and at least one other perspective
• develop and support your ideas with reasoning and examples
• organize your ideas clearly and logically
• communicate your ideas effectively in standard written English

Your essay perspective may be in full agreement with any of those given, in partial agreement, or completely different.

Get more information about preparing for the writing test .

Sample Test Questions

Section 1 - 5 of 30

This action will open a new window. Do you want to proceed?

Welcome to ACT

If you are accessing this site from outside the United States, Puerto Rico, or U.S. Territories, please proceed to the non-U.S. version of our website.

DE Lower Saxony

Recently viewed courses

Recently viewed.

Find Your Dream School

This site uses various technologies, as described in our Privacy Policy, for personalization, measuring website use/performance, and targeted advertising, which may include storing and sharing information about your site visit with third parties. By continuing to use this website you consent to our Privacy Policy and Terms of Use .

   COVID-19 Update: To help students through this crisis, The Princeton Review will continue our "Enroll with Confidence" refund policies. For full details, please click here.

Enter your email to unlock an extra $25 off an SAT or ACT program!

By submitting my email address. i certify that i am 13 years of age or older, agree to recieve marketing email messages from the princeton review, and agree to terms of use., how long is the act.

Time is your enemy on the ACT, and you have to use it wisely. After all, how often do you take a test in school with a minute or less per question? Don’t panic! Get  the ACT timing tips you need to manage your pace throughout each section of the test.

How long does the ACT take?

The ACT is 3 hours long (technically 2 hours and 55 minutes). Including breaks, the exam takes 3 hours and 30 minutes to complete. If you sign up for the optional essay (the ACT Plus Writing ), the test clocks in at 3 hours and 40 minutes or just over 4 hours with breaks.

Here’s how your time is broken up by section:

How many questions are on the ACT?

The ACT has a total of 215 questions with 1 optional essay. The English section has 75 questions, the math section has 60 questions and the reading and science sections both have 40 questions.

What Time Does the ACT Start?

ACT test-takers are required to report to their test center by 8:00 am. The test begins after students are checked in and seated, usually by 8:30 am. Note that if you are taking the test on the computer, you may be taking the ACT at different times, including in the afternoon—be sure to look at your test ticket after you register for the ACT.

Finding the Right Pace for You

Your personal pacing strategy will depend on:

• your  ACT score goals
• what works best for you

See what’s working and what’s not by taking practice tests. We’ve got a free ACT practice test right here. Here are some ACT timing tips to keep in mind as you practice for your test date:

Read More: What's a Good ACT Score?

English Test

The most important thing is to finish. Try finding all the easy questions (the ones you know how to do) on the English Test first. Learn more about working questions out of order in our ACT Tips and Tricks article.

Spend more time to do fewer questions, and you’ll raise your accuracy. How do you know if you’re going too fast? Check your practice tests for careless errors on questions you should have gotten right. Get some ACT math practice here .

Reading Test

Practice extensively to find the pace that works best for you on the Reading Test . 

• Some students are slow but good readers. If you take 35 minutes to do fewer passages, you could get all of the questions right for each passage you do.
• Other students could take hours to work each passage and never get all the questions right. But if you find all the questions you can do on each passage (and guess on the rest of the questions), you could hit your target score.

Science Test

The ACT science test may have either 6 or 7 passages. Be aggressive and keep moving! Spend the time needed to ace the easiest passages first. Then move on the more difficult passages. Even on hard passages, work the the questions that look easiest first.

Read More: How to Write the ACT Essay

Writing Test (the Essay)

Outlining your essay before you write will keep you on track. Plus, organization is key to a great ACT writing score —you'll be doing yourself a big favor!

Remember, there is no guessing penalty on the ACT. Always mark an answer even if you don’t “work” the problem.

Build the right ACT prep plan for you

Our private tutors will help you build a prep plan that's customized to your score goals, study habits, and schedule.

Find a Tutor

Explore Colleges For You

Connect with our featured colleges to find schools that both match your interests and are looking for students like you.

Career Quiz

Take our short quiz to learn which is the right career for you.

Get Started on Athletic Scholarships & Recruiting!

Join athletes who were discovered, recruited & often received scholarships after connecting with NCSA's 42,000 strong network of coaches.

Best 389 Colleges

165,000 students rate everything from their professors to their campus social scene.

SAT Prep Courses

1400+ course, act prep courses, free sat practice test & events,  1-800-2review, free digital sat prep try our self-paced plus program - for free, get a 14 day trial, what would you score on the mcat today.

Thank you! Look for the MCAT Review Guide in your inbox.

I already know my score.

Enrollment Advisor

1-800-2REVIEW (800-273-8439) ext. 1

1-877-LEARN-30

Mon-Fri 9AM-10PM ET

Sat-Sun 9AM-8PM ET

Student Support

1-800-2REVIEW (800-273-8439) ext. 2

Mon-Fri 9AM-9PM ET

Sat-Sun 8:30AM-5PM ET

Partnerships

• Teach or Tutor for Us

College Readiness

International

Advertising

Affiliate/Other

• Enrollment Terms & Conditions
• Accessibility
• Cigna Medical Transparency in Coverage

Register Book

Local Offices: Mon-Fri 9AM-6PM

• SAT Subject Tests

Academic Subjects

• Social Studies

Find the Right College

• College Rankings
• College Advice
• Applying to College
• Financial Aid

School & District Partnerships

• Professional Development
• Advice Articles
• Private Tutoring
• Mobile Apps
• Local Offices
• International Offices
• Work for Us
• Affiliate Program
• Partner with Us
• Advertise with Us
• International Partnerships
• Our Guarantees
• Accessibility – Canada

Privacy Policy | CA Privacy Notice | Do Not Sell or Share My Personal Information | Your Opt-Out Rights | Terms of Use | Site Map

©2024 TPR Education IP Holdings, LLC. All Rights Reserved. The Princeton Review is not affiliated with Princeton University

TPR Education, LLC (doing business as “The Princeton Review”) is controlled by Primavera Holdings Limited, a firm owned by Chinese nationals with a principal place of business in Hong Kong, China.

Choose Your Test

Sat / act prep online guides and tips, (updated) act essay scoring: completely explained.

ACT Writing

It's finally that day you've circled on your calendar—the day when ACT essay scores are released. You log into ACTstudent and look at your essay score . There's an "8" for your overall Writing score as well as four different "domain" scores of 6, 8, 9, and 10. What does your ACT Writing score mean and how is your ACT essay scored? This article will shed some light on both of these things.

Feature image credit: eppny by woodleywonderworks , used under CC BY 2.0 /Resized from original.

A Quick Look Into ACT Essay Scoring

On test day, you complete the first four sections of the ACT then write your essay. What happens next?

Once ACT, Inc. receives your essay, it is scanned and uploaded to an essay grading program for graders to score. In addition, ACT.org states that "[a]n image of your essay will be available to your high school and the colleges to which you have ACT report your scores from that test date."

Each ACT essay is scored by two different graders on a scale of 1-6 across four different domains, for a total score out of 12 in each domain. These domain scores are then averaged into a total score out of 12.

NOTE : The ACT Writing Test from September 2015-June 2016 had a slightly different scoring scale; instead of averaging all the domain scores to get a total ACT Writing score out of 12, the domain scores were combined and scaled into a total score out of 36. On June 28th, 2016, however, ACT, Inc. announced that starting in September of 2016, the Writing test would no longer be scored on a scale of 1-36, due to the confusion this had caused. This change to out-of-12 ACT Writing scores is still different from the pre-September 2015 ACT essay scoring, since that system relied on graders giving the essay one holistic score (rather than 4 analytical domain scores).

Because the ACT Writing is optional, your essay score will not be factored into your ACT composite score. It will, however, be factored into your English-Language Arts subscore, which averages your English, Reading, and Writing scores and rounds up to the nearest whole number.

So what are the four domains that your essay is scored across?

#1: Ideas and Analysis

Scores in this domain relate to your discussion of the perspectives on the essay topic.

#2: Development and Support

Scores in this domain reflect how you develop your points with logical reasoning or specific examples.

#3: Organization

Scores in this domain relate to your essay's organization on both a macro (overall structure) and micro (within each paragraph) level.

#4: Language Use

Scores in this domain depend on your command of standard written English (including grammar and punctuation); variety in sentence structure and vocabulary is also rewarded in this domain.

For more on what goes into each domain score, read my article on the ACT Writing Rubric .

ACT Essay Scoring: Official Policy

Every essay is graded by two graders , who must score the essay within one point of each other. If the graders' scores disagree by more than one point, a third grader will be brought in to resolve the issue. It's currently unclear whether this means a greater-than-one-point difference in domain score or overall essay score between graders—stay tuned for more information.

While your essay receives scores in each of the four domain areas, the domains themselves are graded holistically . For example, in the Language Use domain, there are no guidelines that instruct scorers to deduct 1 point for every 10 grammatical errors.

Another important part of official ACT essay scoring policy is that factual accuracy is not important. ACT essay graders are not supposed to score essays based on whether or not the facts are accurate. The point of the ACT essay is NOT to write a research paper with well-documented facts on a topic. Instead, you're asked to argue in favor of a perspective on the topic and compare your perspective to at least one other perspective ; as long as your examples support your arguments, it doesn't matter if the examples aren't 100% true.

ACT Writing Scores in Practice

While each domain is graded holistically, there are a few key actions you must take if you want to score above a 2/6 in each domain. I've extracted these ACTions via analysis of the essay scoring rubric as well as through scrutiny of the sample essays the ACT provides on its website .

As I go through each domain, I'll be using the following official sample ACT prompt for any examples:

Intelligent Machines

Many of the goods and services we depend on daily are now supplied by intelligent, automated machines rather than human beings. Robots build cars and other goods on assembly lines, where once there were human workers. Many of our phone conversations are now conducted not with people but with sophisticated technologies. We can now buy goods at a variety of stores without the help of a human cashier. Automation is generally seen as a sign of progress, but what is lost when we replace humans with machines? Given the accelerating variety and prevalence of intelligent machines, it is worth examining the implications and meaning of their presence in our lives.

Write a unified, coherent essay about the increasing presence of intelligent machines. In your essay, be sure to

• clearly state your own perspective on the issue and analyze the relationship between your perspective and at least one other perspective
• develop and support your ideas with reasoning and examples
• organize your ideas clearly and logically
• communicate your ideas effectively in standard written English

Your perspective may be in full agreement with any of the others, in partial agreement, or wholly different.

Ideas and Analysis

You must: Have a clear thesis in your essay.

Because you are writing a persuasive essay, it is imperative that you make your position on the topic clear . Otherwise, how can you persuade someone that your view is the correct view?

Since you have limited time and have to compare your perspective with at least one other perspectives anyway, the most straightforward thing to do choose one of the three perspectives given to you by the ACT to argue for in your thesis.

Alternatively, you may choose to take a more nuanced perspective (that may combine aspects of more than one of the ACT perspectives); however, if you do this, it's important to make sure your perspective isn't so nuanced as to be unclear.

For example, acknowledging that intelligent machines challenge what we think humans can be (both in positive and negative ways) is a clear, yet nuanced thesis that combines aspects of Perspectives 1 and 3. Stating that intelligent machines push humans and machines toward new, previously-unimagined possibilities but also take away our humanity because of changes to our daily encounters is an unclear thesis that argues for both Perspectives 1 and 3 being true. This second thesis is not only unclear, but will run into difficulty with the next part of ideas and analysis: comparing your perspective to at least one other.

You must: Discuss the relationship between your perspective and at least one other perspective.

The prompt explicitly states that you need to "analyze the relationship between your perspective and at least one other perspective." If you fail to discuss how your perspective relates to any other perspective, it will be very difficult to score above a 2 or 3 in the Ideas and Analysis Domain.

The easiest way to do this is to take your perspective and compare it to one of the perspectives the ACT lays out for you. With the above "Intelligent Machines" prompt, for instance, you could compare your position to at least one of the following: how machines cause us to lose our own humanity (Perspective One), how they are efficient and create prosperity (Perspective Two), or how machines challenge us and push us to new possibilities (Perspective Three).

It is possible to come up with your own perspective and compare it to another perspective that isn't at all mentioned in the prompts the ACT lays out; however, because time is of the essence in the ACT essay, you'll probably find it more efficient to form your own perspective (possibly from a combination of a couple of the given perspectives) and compare it to the ACT perspectives that don't match yours.

Development and Support

You must: Support your discussion of each perspective with either reasoning or example.

There are a couple of ways you can support your arguments. One way is to use reasoning , which tends to be more abstract. For example, if you were using reasoning to support your argument for Perspective Two, you could discuss how machines taking over lower skill jobs frees up humans to do higher skilled tasks that require more creative thinking.

The other way you can support your points is through use of specific examples . For example, to support Perspective Two, you could use the example of how the mass-production of clothes has made it less expensive for everyone to own things like good boots.

For a high score in this domain, you must: Discuss both positive and negative aspects of the perspectives you disagree with as well.

In order to achieve a high score in this domain, you must show that you understand the complexities of the issue. The main way to do this is to discuss the pros as well as the cons of the perspectives you disagree with.

For instance, if you agree with Perspective Two in the above prompt (machines make us more efficient and that's good), if you compare it to Perspective One, you should provide a brief instance of that perspective being "sort of" true before moving on to show how it is not as true as Perspective Two. Learn how to juggle both sides of a perspective in our article on how to write an ACT essay step-by-step .

Organization

You must: Group your ideas logically.

Writing an organized essay will make it easier for the essay graders to follow your logic and reasoning. Grouping your ideas logically can mean separating out ideas into different paragraphs (for instance, putting each perspective into its own paragraph), or it can involve clearly linking different aspects of the same idea in the same paragraph . No matter how you plan out your essay, try to make it as easy as possible to follow your arguments.

Language Use

You must: Write clearly.

Being able to communicate clearly is a key skill for college and life in general, so it makes sense that it would be tested on the ACT (a college entrance exam). ACT essay graders care more about the clarity of your thoughts than the fanciness of your language. Clarity of writing normally entails using proper grammar and clear, non-convoluted sentence structures. Throwing in fancy vocab won't get you anywhere if it makes things less clear instead of more clear (I've seen this happen too many times to count).

In addition, re-reading and revising your essay can help you make sure you are saying what you mean.

Example of an unclear sentence: Machines are more practical because they are cheaper and so you can hire less people to do the work and pay less money overall and so you have a better profit margin.

Example of a clearer sentence (revised): Machines are more practical and cheaper in the long run because you can hire fewer people to get the same work done.

TURNS out, the steam engine was more practical (and cheaper in the long run) than a thousand people pushing and pulling a train by hand.

What Does This Mean for Your ACT Essay?

From the lists of actions above, you can probably tell that the most important part of the ACT essay is to be clear . The ACT Writing test is designed to measure insight, not just how advanced your vocabulary is. Remember to...

#1: Be clear up front what your perspective on the issue is. Don't hide your thesis.

#2: Make it obvious when you're discussing your perspective (and make sure to discuss the relationship between your perspective and at least one other ).

#3: Support each argument with reasoning and/or specific examples.

#4: Take time to plan so you can write an organized essay.

#5: Focus on writing clearly before you start worrying about using advanced vocabulary.

What's Next?

Want to learn more about how to write an ACT essay? Read my step-by-step guide to ACT Writing .

You've learned what your essay needs to include. But how you do you decode the prompt? Follow along as I teach you how to attack ACT Writing prompts .

Is a longer ACT essay always a better ACT essay? Find out how essay length can affect your score on ACT Writing here .

Want to improve your ACT score by 4 points?

Check out our best-in-class online ACT prep program . We guarantee your money back if you don't improve your ACT score by 4 points or more.

Our program is entirely online, and it customizes what you study to your strengths and weaknesses . If you liked this ACT Writing lesson, you'll love our program. Along with more detailed lessons, you'll get your ACT essays hand-graded by a master instructor who will give you customized feedback on how you can improve. We'll also give you a step-by-step program to follow so you'll never be confused about what to study next.

Check out our 5-day free trial:

Laura graduated magna cum laude from Wellesley College with a BA in Music and Psychology, and earned a Master's degree in Composition from the Longy School of Music of Bard College. She scored 99 percentile scores on the SAT and GRE and loves advising students on how to excel in high school.

Student and Parent Forum

Our new student and parent forum, at ExpertHub.PrepScholar.com , allow you to interact with your peers and the PrepScholar staff. See how other students and parents are navigating high school, college, and the college admissions process. Ask questions; get answers.

Ask a Question Below

Have any questions about this article or other topics? Ask below and we'll reply!

Improve With Our Famous Guides

• For All Students

The 5 Strategies You Must Be Using to Improve 160+ SAT Points

How to Get a Perfect 1600, by a Perfect Scorer

Series: How to Get 800 on Each SAT Section:

Score 800 on SAT Math

Score 800 on SAT Reading

Score 800 on SAT Writing

Series: How to Get to 600 on Each SAT Section:

Score 600 on SAT Math

Score 600 on SAT Reading

Score 600 on SAT Writing

Free Complete Official SAT Practice Tests

What SAT Target Score Should You Be Aiming For?

15 Strategies to Improve Your SAT Essay

The 5 Strategies You Must Be Using to Improve 4+ ACT Points

How to Get a Perfect 36 ACT, by a Perfect Scorer

Series: How to Get 36 on Each ACT Section:

36 on ACT English

36 on ACT Math

36 on ACT Reading

36 on ACT Science

Series: How to Get to 24 on Each ACT Section:

24 on ACT English

24 on ACT Math

24 on ACT Reading

24 on ACT Science

What ACT target score should you be aiming for?

ACT Vocabulary You Must Know

ACT Writing: 15 Tips to Raise Your Essay Score

How to Get Into Harvard and the Ivy League

How to Get a Perfect 4.0 GPA

How to Write an Amazing College Essay

What Exactly Are Colleges Looking For?

Is the ACT easier than the SAT? A Comprehensive Guide

Should you retake your SAT or ACT?

When should you take the SAT or ACT?

Stay Informed

Get the latest articles and test prep tips!

Looking for Graduate School Test Prep?

Check out our top-rated graduate blogs here:

GRE Online Prep Blog

GMAT Online Prep Blog

TOEFL Online Prep Blog

Holly R. "I am absolutely overjoyed and cannot thank you enough for helping me!”

How Does the ACT Work, and What’s a Good Score?

[et_pb_section bb_built=”1″][et_pb_row][et_pb_column type=”4_4″][et_pb_text _builder_version=”3.10″]

The ACT is one of the two main college readiness exams. Like they do with the Scholastic Aptitude Test (SAT), admissions offices use your ACT score to help determine if you’re a good fit for their college or university.

Both tests are widely used, but recently, the ACT has become more popular than the SAT. Over two million students take the ACT every year, while about 1.7 million take the SAT. It’s easy to feel like all standardized tests are the same, but these tests measure two different things.

The SAT measures your ability to learn, and the ACT measures what you’ve already learned. They’re two completely different approaches to determining your “college readiness,” but both are well-respected, and most schools don’t have a preference for one or the other.

The ACT was developed by a college professor in response to his frustrations with the SAT, which at the time was the only means of gauging how prepared students were for college. As an educator, he felt it was far better to know that your students had the necessary prerequisite knowledge than to know how good they were at learning.

Tomato, to-mah-to, you’re taking the ACT. Here’s what you need to know about it.

Quick facts on the ACT

Eligibility requirements: none

Length: three hours and 30 minutes

Sections: four required (English, math, reading, science), one optional (writing)

Number of questions: 215 (plus an essay prompt)

Types of questions: multiple choice, essay (writing section only)

Score range: 1–36

Cost: $46 without the writing section, $62.50 with the writing section

What does ACT stand for? ACT technically stands for American College Testing, but the test is administered by an organization of the same name, and they simply refer to it as the ACT.

What does the ACT test you on?

The ACT tests you based on things you should learn by the end of your senior year of high school. We’ll tell you what to expect from each section and give you some tips, but first, here’s a tip that applies to every multiple choice section:

Never leave a question blank. You aren’t penalized for wrong answers, so even if you have absolutely no clue what the answer is, take a guess!

ACT English Test

Length: 45 minutes

Number of questions: 75

The English section of the ACT will measure your mastery of the nuts and bolts of prose. Rather than asking you to retrieve information or think critically about a passage, you’ll examine individual phrases, sentences, or paragraphs.

The makers of the test say that the English section “measures your understanding of the conventions of standard English (punctuation, usage, and sentence structure), production of writing (topic development, organization, unity, and cohesion), and knowledge of language (word choice, style, and tone).”

There are several varieties of multiple choice questions, but they’re all based on passages. If a passage has underlined sections, some of the questions will ask you to improve those parts with one of the options provided. If none of the suggested changes improve the passage, you can choose “no change” (this is pretty common).

You may be asked to change a passage to create a specific effect, like building a stronger transition between two thoughts or rearranging sentences to form a more logical progression. You’ll also be asked about how particular changes might affect the passage, or what information a particular section provides.

According to national data from the ACT, this is the section most people score lowest on.

Tip: This test gives you the least amount of time per question, so don’t waste time on tough questions. Guess, and come back to them later if there’s time.If you don’t have extra time at the end, think about how many more questions you were able to get through!   It’s sunk-cost economics at work.

ACT Mathematics Test

Length: 60 minutes

Number of questions: 60

All of the questions of the math section of the ACT are solvable without a calculator. Still, using one could save you some precious time. Here’s how the makers of the test break down the questions:

• Number and quantity (7–10%)
• Algebra (12–15%)
• Functions (12–15%)
• Geometry (12–15%)
• Statistics and probability (8–12%)
• Modeling (less than 25%)
• Integrating essential skills (about 15%)

You’ll need to know the basic terminology and formulas for basic high school math. Be sure to read how the ACT describes each math section .

The ACT has already done a huge favor for you: the questions in each section appear in order of difficulty, so the easiest ones come first. That doesn’t always mean they’ll be the easiest ones for you , so you’ll still want to skim ahead if you get stuck on one.

Tip: Before you start solving a problem, take a look at the possible answers. They may help you decide how to approach the problem.

ACT Reading Test

Length: 35 minutes

Number of questions: 40

The reading section of the ACT contains four passages with 10 questions each. These are designed to test your reading comprehension, and the questions will ask you to summarize pieces of the passage, make inferences from the text, and identify key information. The bulk of the questions will be about the ideas and details in the passages.

The four passages come from the same four subjects every year, and they always appear in the following order:

• Social science
• Natural science

According to national data from the ACT, this is the section people score highest on.

Tip: Answers on the reading section are deliberately trying to trick you. There may be multiple answers that look correct. Before you read the possible answers, try to answer the question for yourself when possible. Then it’ll be harder for those almost right answers to mislead you—because you’ll be looking for a specific answer.

ACT Science Test

For the science section of the ACT, you won’t be expected to have advanced knowledge of any particular kind of science, but you will need a general understanding of these subjects:

• Meteorology

The majority of this test involves interpreting data and working with experiment results. Like the reading and English sections of the ACT, science questions are based on a passage. You’ll have between four and seven questions per passage.

The passages you’ll work from will take these three formats:

• Data representation (30-40%)
• Research summaries (45-55%)
• Conflicting viewpoints (15-20%)

Data representation questions will focus on tables and graphs, and they’ll test your ability to interpret scientific data.

Research summaries will focus on experiments, and you’ll be asked about the design of the experiment and the conclusions drawn from the results.

Conflicting viewpoints questions ask you to analyze and compare two hypotheses or views that are based on different premises or incomplete data.

ACT Writing Test

Length: 40 minutes

Number of questions: one

The writing section is the only part of the test that doesn’t affect your composite score. (It is optional, after all.) The test will present three perspectives on an issue, and you’ll respond with an essay.

Regardless of the prompt, your essay will have the same three expectations:

• Analyze and evaluate the perspectives given.
• State and develop your own perspective on the issue.
• Explain the relationship between your perspective and those given.

Should I take the ACT Writing Test?

It definitely won’t count against you if you don’t take the optional section of the ACT. So if you’re a terrible writer, that’s probably a good choice. However, since it doesn’t count towards your composite score, and it can provide the admissions office with a more personal representation of your academic capabilities, it’s a good idea to take it.

A big shiny ACT score can certainly make your college applications stand out, but people can only learn so much about you from a number. There’s no guarantee a college will use your ACT essay, but if you’re a good writer, this is a positive differentiator that not every applicant is going to have.

Personally, I’d recommend it.

How do ACT scores work?

On the ACT, you gain points for correct answers, and you don’t lose points for incorrect answers.

You’ll receive a separate score for each section of the ACT: English, math, reading, science, and writing (optional). Scores range from 1 to 36 on each section—with the exception of the writing section, where scores range from 2 to 12. These scores are averaged into your “composite score.”

Your score report will show your score and percentile in each section, plus two additional sections: STEM , and ELA (English Language Arts). These are good indicators of your competency in specific subjects, but when someone says, “What’s your ACT score?” they’re asking about your composite score. ACT.org can walk you through each part of your ACT score report .

You’ll be able to view your scores online about two weeks after the exam, and you’ll get your score report in the mail within 2-8 weeks.

What’s a good ACT score?

While a perfect SAT score is 2,400, a perfect ACT score is just 36. But less than .1% of all ACT test-takers get a perfect score. So what’s a good ACT score?

The average ACT score is 21. But sitting smack dab in the middle isn’t going to make your application stand out. And for schools that want to build on their reputation for excellence, a mediocre score may not be good enough (unless you have an exceptional portfolio or something else to show off).

A good score is different for every school

Unless a school explicitly states a minimum ACT score you need to be admitted, you have to look at the students who are actually accepted to determine a “good score.” Whatever schools you’re looking at, you can see the range of standardized test scores for students they’ve accepted. You can even compare the bottom 25% of accepted students and the top 75%. If the other pieces of your application are strong, you can probably get by with a lower score.

Ivy League schools , for example, have significantly higher standards than the national average, and their students’ average ACT scores range from as low as 29 to as high as 35 (or from the 92nd percentile to the 99th).

So a good score really depends on what school you want to go to. But unless you’re looking at a top-tier school—one with a notoriously low acceptance rate—scoring in the top 75% of all test takers should be good enough for most schools. That means a good score is 24 or higher.

But what do you mean by “a good score”?

Of course, good is a relative term. Maybe by “good” you mean scoring in the top 90%? In that case, you’re shooting for a 28 or higher. And yes, four points can make that big of a difference. Sometimes getting a single point higher on your composite score can make a difference of 7%.

While a score around 24 should get you into most schools, scoring 28 or higher will have some schools throwing money at you, begging you to be one of their students.

Now that you know more about what to aim for, let’s talk about what to do with a bad score.

What happens if you get a bad score on the ACT?

No matter how much you study, there are lots of factors that can add up to a bad test day. Lack of sleep, poor diet, not exercising regularly, health issues, stress, and distractions make it harder to perform your best under pressure.

Thankfully, you can retake the ACT as many times as you want. And ACT.org says that 57% of people who retake the ACT improve their score.

When it comes time to submit your ACT score to a school, you can choose to submit your highest composite score. So if you have a bad test day, no one even has to know. This doesn’t mean you can pick and choose your best results from each section, though. If you did really well on different sections on different days, you may want to share both scores—but you can’t blend them into a new composite score like some kind of ACT Frankenstein.

How do you prepare for the ACT?

Like all standardized tests, the best way to prepare for the ACT is to take practice tests. You should certainly study the terms and concepts you’ll need to be familiar with (especially for the math and science sections), but when it comes down to it, practice is the best route to a high score.

The best practice tests will give you explanations for why one answer is correct, and the others are wrong. To make sure you’re getting the most relevant practice, be sure to check out the materials provided by ACT.org, the makers of the test.

Here are some free practice questions they provide, with explanations for each answer:

• English practice questions
• Math practice questions
• Reading practice questions
• Science practice questions
• Writing samples

Let us know in the comments if there’s anything else you wish we would’ve told you about the ACT. Good luck in your studies!

[/et_pb_text][/et_pb_column][/et_pb_row][/et_pb_section]

Spread the word:

You might also like

Does ACT cater for students with any form of learning disability? For instance this student is able to write ACT but would need special accommodation, like additional time, would such special accommodation be allowed for the ACT exam? This student has a condition called Impaired Processing Speed (IPS). An official report to that effect can always be forwarded to the relevant office of ACT when and if necessary.

I kindly await your advice.

Yes, the ACT offers accomodations for documented learning needs. Here’s where they talk about that: https://www.act.org/content/act/en/products-and-services/the-act/registration/accommodations.html

Leave a Comment Cancel reply

Your email address will not be published. Required fields are marked *

Save my name, email, and website in this browser for the next time I comment.

Notify me of follow-up comments by email.

Notify me of new posts by email.

Privacy Overview

• March 2, 2018

How Long is the ACT?

For those with the question “how long is the ACT,” you are already thinking ahead to one of the biggest constraints of the ACT : time. With a minute per question, being able to answer questions efficiently and effectively is key!

Many choose to think of the ACT as more of a marathon than a sprint. It’s important to know how to pace yourself on the test as well as take advantage of breaks to recover before jumping back into the race. Knowing exactly what you are getting yourself into and being prepared for the test is a great way to avoid any unnecessary stress or concern on test day.

So read on and see how long the ACT takes and what exactly to expect.

How Long Does the ACT Take?

• The ACT is 2 hours and 55 minutes in length of actual sit-down testing. Including breaks, the ACT is 3 hours and 35 minutes long .
• If you take ACT with optional essay, the testing time is 3 hours and 35 minutes . Including breaks, the ACT with optional essay is 3 hours and 50 minutes long.

ACT Test Length By Section:

• Section 1 – English – 45 minutes
• Section 2 – Math – 60 minutes
• Section 3 – Reading – 35 minutes
• Section 4 – Science – 35 minutes
• Optional Writing – 40 minutes

The testing usually begins between 8:30AM and 9:30AM, although all students are required to arrive at the test-taking center at 8AM for check-in. Let’s have a look at the schedule broken out by section for an 8:30AM exact start time:

[table id=10 /]

Generally speaking, this is the schedule you will experience across all testing centers. Sometimes there are differences in this schedule based on the strictness of the proctor on break time or if there is any kind of problem or disruption during the test that needs to be taken care of.

After the Math section of the MCAT, all test takers receive a 10-minute break. If your proctor attempts to skip this break, please let them know that you would like this time for a rest, as you are entitled to this 10-minute break to regroup.

What Else You Need to Know About Testing Day

Students must be checked in by 8AM Saturday morning of your test date, so it’s recommended to show up at the testing center at least 15 minutes beforehand. It’s important to keep ahold of your admissions ticket even after check-in, as you will need some of the information when filling out personal details on the test (check out our ACT checklist here! ).

The variation occurs in how long it takes to check everyone in, which is why the start time can vary between 8:30 and 9:30 AM. Once the proctor has checked everyone in and read the directions, you will then need to fill out all the personal details as directed. Assuming all goes well, you then will take the test and expect to be finished between 11:30AM and 1PM.

• TERMS & DISCLOSURE
• PRIVACY POLICY

BLACK FRIDAY & CYBER MONDAY SAVINGS

Discounted mcat, gre, lsat, nclex, sat, act.

How Long Does The ACT Take? An Overview of the ACT Test Length

The ACT—you know it, you love (to hate) it, and you need to know more about it in order to prepare for it effectively. Even with the explosion in the popularity of test-optional admissions policies that occurred during the COVID pandemic, getting a good score on the ACT can still help your college application, which leaves many students in the position of wanting or needing to take the test.

If that applies to you, then it’s important to know that the first step to doing well on the ACT is understanding as much as possible about the test. To that end, your friendly neighborhood test expert is here to answer one of the most common questions from students: how long is the ACT? We’ll cover the length of individual sections and the entire test, as well as discuss some strategies for coping with the pace of the test. Let’s dive in.

Table of Contents / Quick Reference Guide:

An overview of the act exam.

• How Long Does The ACT Take?
• ACT Test Sections & Time
• How Many Questions Are On The ACT?
• ACT Start & End Times
• Extended Time Accommodations for the ACT
• Time Management Tips For The ACT
• Frequently Asked Questions About The Length of the ACT

The ACT is one of two tests—with the other being the SAT—that are used by colleges and universities in the admissions process to assess students’ readiness for college-level coursework. As I alluded to above, many schools implemented a test-optional admissions policy during the COVID pandemic, and most of those schools have indicated that they will remain test-optional for at least the foreseeable future. Test-optional is not the same thing as test-blind , however: schools with a test-optional policy will still consider test scores if you submit them with your application, which means that scoring well on the ACT can be a major boost to your chances of being admitted to one of those schools. There are also test score requirements attached to some scholarships, both athletic and academic. All of this means that the ACT still occupies an important place in the admissions landscape, which makes preparing for it a part of the high school experience for a large number of students.

Part of preparing for the test is knowing your enemy, from high-level strategy to content all the way down to logistical details. That leads us nicely to the main point of this post: tackling the question of ‘how long is the ACT?’

How Long Does The ACT Take To Complete?

In terms of purely testing time, the length of the ACT taken with standard time clocks in at 175 minutes—or just under 3 hours —without the Writing section; if you’re taking the Writing section, your total testing time will be 215 minutes, or just over 3.5 hours . You’ll use this time to complete 215 multiple-choice questions distributed between English, Math, Science, and Reading, as well as the optional Writing section if relevant.

Of course, there are other aspects of test day besides just the time you spend on each section. The proctor will read out instructions at the beginning and end of the test, and you’ll spend some time filling out biographical information on your answer sheet; additionally, a break of roughly 10 minutes is incorporated into the test after the Math section, as is a shorter break after Science if you are taking the Writing section. Finally, you may sometimes need to complete another short multiple-choice section at the end of the test—this is an experimental section that allows the ACT to test out questions for future administrations, and it does not affect your score.

Related Reading: What’s a Good ACT Score for 2023 & 2024?

With all of that factored in, the ACT’s website recommends that students testing with standard time should plan to be at the test center from 8:00am until approximately 12:35pm if they are not taking the Writing section or from 8:00am until approximately 1:35pm if they are taking the Writing section .

How Long Does Each Section of the ACT Take?

Though it’s helpful to know how long you should expect to be at the test center on the day you take the ACT, it’s more important as far as your preparation goes that you know how that time is distributed between the sections. The image below, taken from the ACT’s website, provides a high-level summary of the timing for each part of the test. For quick reference, the English Section of the ACT is 45 minutes long, the Mathematics section of the ACT is 60 minutes long, the Reading section of the ACT is 35 minutes long, and the Science section of the ACT is 35 minutes long. Finally, the optional Writing section of the ACT is 40 minutes long.

Source: ACT.org

ACT Test Times & Sections

If you’re looking for a more in-depth breakdown of the timing for each section, take a gander at the table I’ve created below for additional details.

If you’re looking for information about taking the ACT with extended time , don’t worry—we’ve got a whole section for you later in the post. Keep reading for additional details.

How Many Questions Are On the ACT?

The ACT exam has a total of 215 questions. These questions are broken down into four main sections we covered above: English, Math, Reading, and Science. The number of questions for each section varies, with the English section having the fewest and the Math section having the most. The English section has 75 multiple-choice questions, while the Math section has 60 multiple-choice questions. The Reading section has 40 questions, and the Science section has 40 questions as well.

Want To Get The Highest Score Possible On the ACT?

Act start and end times.

It will be pretty clear to you from the information above that the ACT is not a short test, even if you’re not taking it with extended time. The ACT takes about a minimum of about 3 hours , and you should expect to be at the test center for noticeably longer than that, especially if you’re taking the Writing section.

As I mentioned earlier, the ACT recommends that you arrive at your test center no later than 8:00am . Once you arrive, you’ll spend a few minutes checking in, finding your room, and getting situated; then, the proctor will begin the testing process by guiding you through the pre-test procedures. The specific amount of time required for this part of the process will vary depending on the proctor and the group of students; however, you can expect that you’ll likely begin taking the actual test somewhere between 8:30am and 9:00am.

One interesting aspect of ACT timing is that the first “half” of the test is significantly longer than the second half: English and Math combine for 135 minutes of standard-time testing, whereas Reading and Science comprise only 80 minutes. The upshot of this is that by the time you take your first break after the end of the Math section, you’re likely to be pretty fatigued; however, it’s important to remind yourself that you’re already close to two-thirds of the way through the test (if you’re not taking the Writing section). Use the break to rehydrate and refuel for the home stretch.

Once you finish the second half of the test, the proctor(s) will collect your materials and read a couple of final instructions; at that point, you’ll finally be free. Students testing with standard time are typically released from their testing room at roughly 12:35pm if they’re not taking the Writing section and approximately 1:35pm if they are taking the Writing section.

Extended Time Accommodations For the ACT

There are a number of accommodations available for students on the ACT, but one of the most common is extra testing time, or extended time (ET). This accommodation can take slightly different forms depending on the student, as the ACT tries to approve students for testing accommodations based on the accommodations that they receive in school , but probably the two most common versions of ET are 1.5x extended time (“time and a half”) and 2x extended time (“double time”).

Previously, ET on the ACT was administered as one lump amount: students would be given 1.5x or 2x the total amount of testing time and allowed to distribute it among the sections of the test as they saw fit. This is no longer the case, however. Instead, the ACT has adopted a more intuitive approach to ET wherein students are simply given 1.5x or 2x the standard time limit to complete each individual section. See the table below for specific numbers.

The ACT is a long test to begin with, but it can be an especially grueling experience for students who have extended time. Because of this, many students who are approved for the double time accommodation may also be granted the option to test over multiple days, which makes the testing experience somewhat more manageable.

If you use accommodations in school or think you may be eligible for accommodations on the ACT for another reason, such as a learning disability or recent injury, contact your school counselor for more information on how to apply for testing accommodations. You can also check out the ACT’s website for additional details about ET and other accommodations.

Time Management Tips for the ACT Exam

It goes without saying that knowing how long the ACT is will only be your first step to succeeding on the test; the next step is making sure that you’re able to complete each section within the time limit. Here are a couple of tips for managing your time on each section of the test.

ACT English Time Management Tips

• Read (or at least skim) the passage as you go. It may seem counterintuitive that doing more reading will improve your timing, which is why many students instinctively skip over the portions of the passage that don’t have any questions attached to them. There are two things you may not realize, however. First of all, context is your friend on many of the grammar questions, especially those annoying transition questions—knowing what came right before or comes right after after the sentence being asked about will make these questions significantly easier, and therefore faster. And second, most passages have 1-2 questions at the end that ask about the content of the entire text. You don’t need to read deeply enough to memorize specific details in order to answer these end-of-passage questions, but having a general sense of what the passage is about will help you address them more quickly.
• Use process of elimination (POE). This is true across the entire test, but English is one of the sections where POE can be most useful. Look for opportunities to play answer choices against each other. For example, semicolons and periods play the exact same role on the ACT 99.8% of the time—if you see two choices that are identical other than a period being swapped out for a semicolon, you can almost certainly safely eliminate both of those answers and focus on the others.

ACT Math Time Management Tips

• Break the section up into thirds. As I alluded to earlier, you shouldn’t expect every problem in Math to take you the same amount of time, despite the fact that you have a nice, round 1 minute per question in this section. Rather than approaching the Math section as one big blob of 60 questions, think of it as three subsections: #1-20 (the easiest portion), #21-40 (the medium-est portion), and #41-60 (the hardest portion). Give yourself a time goal for each of these subsections; then, as you practice, work to fine-tune those goals in order to give yourself as much time as possible for the harder problems while not moving so fast that you’re overly prone to careless errors on the easier problems.
• Prioritize, prioritize, prioritize. The range of content that can show up on an ACT Math section is massive, and very few students can go into the section expecting to be comfortable with absolutely every concept that could show up. As you get further into the section and the questions get harder, you should grow commensurately more willing to skip past questions that test content you’re not familiar with and prioritize those where you’re more comfortable. Remember that every question in this section counts exactly the same—if you’re going to run short on time, make sure the questions you have to guess on are the ones that you would have had the hardest time getting right anyway.
• Use your calculator! This one should be pretty obvious, but nonetheless students frequently forget how helpful of a tool their graphing calculator can be. In addition to using the calculator for complex arithmetic, look for opportunities to save yourself time by doing things like graphing equations.

ACT Reading Time Management Tips

• Address the section in the way that makes the most sense for you. Remember that the passages will always be the same four genres in the same sequence . If you notice during your practice that you are consistently performing worse on one specific passage-type, consider saving that passage for last in order to ensure that you have the time to bank the points on the passages that you find easier. Similarly, remember that you can answer the questions within an individual passage in whatever order is easiest for you. For most students, it’s better to answer the questions that ask about a specific portion of the passage first; in the process of close-reading the text to address those problems, you’ll naturally grow your understanding of the passage and therefore make the broader questions significantly easier and faster.
• Watch out for the ‘EXCEPT’ questions. Probably the question archetype that has the greatest potential to bog students down is the ‘EXCEPT’ question. Rather than asking you which of the four answer choices is correct, these questions ask you to identify the choice that is NOT found in the passage. To answer these monsters, you’ll need to locate the other three answer choices in the passage and use POE to identify the one that isn’t present, which is extremely difficult to do quickly. Save these questions for last, and be conscious of the clock—you may need to cut bait, take your best guess, and move on to the next passage for the sake of time.

ACT Science Time Management Tips

• Don’t read unless you need to… The dirty secret of this section is that most of the text will simply give you information that you could have gleaned more quickly from the graphs and charts. When you first start a passage, take a quick look at the figures to get the lay of the land, then go straight to the questions. Remember that you can always go back to the passage to read and get more information if questions require it, which they sometimes will; however, you can’t get back the time that you spend reading stuff that isn’t tested.
• …and if you do need to, don’t read the way the test wants you to. One of the six passages in this section—the Conflicting Viewpoints passage—will have few if any visual aids. There’s no getting around it: you’ll need to actually do some reading in this passage. The best way to approach this passage-type, however, is not to read the whole passage and then answer the questions in order. Instead, read the first portion, then answer any questions (and do any POE) that you can using that information; then, go to the next portion and do the same thing, and rinse and repeat until you’ve completed all of the questions. This allows you to keep less information in your brain at one time, and it also ensures that you’re earning points along the way rather than front-loading all of the time-consuming reading and not answering questions until the end, when you’re likely to be more frazzled and hurried.

Closing Thoughts On The Length of the ACT

The primary challenge of the ACT for most students is its timing, and understanding the specifics of that timing is the first step to mastering the test. Once you have a grasp on the information I’ve provided in this post, the next step is to incorporate it into your practice. Make sure you’re giving yourself plenty of time to prep for the ACT—most students require a significant amount of work and many practice reps before they start to feel comfortable with the speed of this test. You may feel overwhelmed at first by how fast everything goes, but don’t get discouraged: with practice, you can master the pace of the ACT and raise your score.

Frequently Asked Questions About The ACT’s Length

What happens if i’m late to the act.

It’s crucial to be aware of the ACT guidelines when it comes to test day logistics. The organizers are quite strict about latecomers and may deny entrance if you arrive after the designated time. To ensure a smooth experience and guarantee admission, I recommend getting to the test center between 7:30 and 7:45 am. Typically, you’ll find a line of other students waiting outside, so it’s best to arrive early. The doors to the test center open at 7:45 am and give you a 15-minute window of opportunity to enter before they close promptly at 8 am. Once the doors shut, late entry is not permitted, so arriving on time is absolutely essential.

How long is the ACT without the Writing section?

The ACT Writing section, or Essay, is an optional portion of the test. Because colleges and universities don’t typically put a ton of emphasis on it during the application process, many students opt not to take the Writing section. If you take the ACT without Writing, your total (standard) testing time will be 175 minutes , and the ACT’s website states that you’ll be released from the testing center at approximately 12:35pm.

How many questions can you miss and still get a 36 on the ACT?

This is a very difficult question to answer with certainty: the scaled score conversion tables vary depending on the difficulty level of each section, which means that you’ll have more wiggle room to get questions wrong if your section is harder than average. Speaking broadly, though, we can say that to get a 36 in English or Math, you typically can’t miss more than 2-3 questions; to be guaranteed a 36 in Reading or Science, you have to have perfect accuracy. As a result, to be confident of getting a 36 on the ACT, you’re probably aiming to miss a total of 5-6 questions across the entire test, with most if not all of those needing to be concentrated in English and Math.

It’s also important to remember, however, that your section scores are averaged to calculate your composite score; as a result, you don’t need to get a 36 on every individual section to score a 36 on the ACT as a whole. You can get a composite score of 36 by earning 36s in two sections and 35s in the others, or even 36s in three sections and a 34 in the fourth. This gives students verrrrrry slightly more margin for error when it comes to earning a perfect score on the ACT.

Which is longer, the SAT or the ACT?

Currently, the SAT is almost exactly the same length as the ACT without Writing: the ACT without Writing clocks in at 175 minutes of testing time , while the SAT (without its Essay, which is only available in specific locations) comes in at 180 minutes of testing time. This means that the lengths of the two tests shouldn’t currently be a factor in your decision of whether to take the ACT or the SAT .

In March of 2024, however, the SAT will be switching to its new digital form for US testers. Once that happens, the SAT will be shortened to 134 minutes of testing time , which will make it a noticeably shorter test than the ACT. Test length is still unlikely to be the deciding factor between the two tests for most students, but it will become more of a consideration once this shift occurs.

Related ACT Resources:

How Much Does The ACT Cost?

Should You Take The ACT With or Without Writing?

What’s a Good ACT Score For Ivy League Colleges?

ACT Test Dates For 2023 & 2024

What’s a Good ACT Score For 2023 & 2024?

Understanding the Differences Between the ACT & SAT

• September 27, 2023
• Standardized Tests & Test Prep

Like this article?

Recent posts, act and sat score comparison: score conversions and chart.

ACT Test Dates 2024: Upcoming Test Dates & Deadlines for the ACT Exam

How To Check Your SAT Score: A Step-By-Step Guide

How to improve your act score: 16 expert strategies.

SAT Test Dates 2024: Upcoming SAT Dates & Registration

How Many Times Can You Take The SAT Exam?

Sign Up For the Latest News

• 60 Chelsea Piers Ste 6020, New York, NY 10011
• [email protected]
• 844 663 9484

An employment lawyer explains the 3 things you should do the moment you're put on a PIP

• Craig Levey is an employment law attorney of over 12 years.
• He says that PIPs are often used by companies to "paper the file"  once they've already decided to fire the employee. 
• If put on a PIP, you should read it carefully, consider if it's reasonable, and decide how to respond.

This as-told-to essay is based on a conversation with Craig Levey, an employment law attorney and partner at Bennett & Belfort, P.C., a law firm based in Cambridge, Massachusetts. The following has been edited for length and clarity.

In my experience as an employment law attorney of 12 years, performance-improvement plans, or PIPs, result in termination in most cases. While there are circumstances where an employee survives the PIP and keeps working there, those situations are rare.

Many companies use PIPs to paper the file; they document the areas in which this employee is deficient, and then once the PIP is over — or sometimes even sooner — they terminate the employee and say, "As we outlined in the PIP, things weren't going well in certain areas. We don't feel like you've improved, so we're going to part ways with you."

PIPs are a tool for companies to prevent litigation, too, because then they can argue that they didn't terminate the employee because of discrimination, retaliation, or sexual harassment but rather because of performance.

When an employee receives a PIP , there are three steps that I suggest that they follow.

1. Read the PIP very carefully

Many people are so shocked that they receive a PIP that they don't actually take the time to take a deep breath, sit down, and read it.

You want to fully understand the language that's in the PIP and what the expectations are. If you need clarification because there's not enough information, you should go back and ask your company your questions and see if they offer support, such as one-on-one meetings with your supervisor.

2. Decide if the PIP is reasonable

Once you've read the PIP, ask yourself, is this PIP designed for me to fail, or is it reasonable?

A good PIP should be very clear about the areas where the employee is supposedly deficient. In a reasonable timeframe, it should spell out how that employee will get better and offer support such as regular meetings with their supervisor.

A bad PIP is very short and doesn't articulate the areas in which the employee needs to get better, so the employee is left confused and doesn't know what is being asked of them. There either isn't a timeframe given, or the timeframe is unreasonable, and no support is offered to help the employee improve.

Related stories

After receiving a bad PIP, the employee walks away confused or feeling unreasonable. Typically, when an employee receives a bad PIP, it means the company doesn't want them to improve. It's just a tool to pave the way for the termination.

3. Decide whether to respond

If, after reading your PIP, you feel it's actually pretty reasonable and you can satisfy the requirements in the time allotted, you don't need to respond. Just move on and focus on working on what's needed and satisfying the requirements laid out.

But sometimes, there are unfair PIPs with requirements that even the best employee in the world could never satisfy. If the PIP is unrealistic or unreasonable, then you need to respond.

If you decide to respond, email your supervisor or HR and spell out why it is unreasonable. Perhaps the timeframe is not long enough, or the expectations are unattainable, or the company isn't offering any support.

By spelling out the ways the PIP is unreasonable, it can protect your interests a bit in the event that you are later terminated.

Is the PIP retaliatory?

You should also think about whether the PIP is retaliatory. A lot of companies will issue a PIP in retaliation for the employee complaining about something illegal going on in the workplace or taking leave for disabilities or pregnancy, which is protected under the Family and Medical Leave Act (FMLA) .

Leave is there for a reason. If you have a disability, if you have mental health issues, or if you're pregnant, then you need to use your leave options because that's precisely what they're there for.

Unfortunately, I see all the time that companies perceive employees who take leave to be a burden. Companies don't like when they have somebody out of work and they have to satisfy those duties, either by hiring somebody to fill in or asking somebody else to take on those duties.

This happens far too often across the US. When you look at other countries around the world — how they value leave and how much time off they give people — there's more emphasis on quality of life rather than the American way, which is focused on the "grind." American companies put so much more emphasis on the bottom line and making money and less on employee well-being.

If you believe your PIP is due to your company retaliating against you, then you can respond in writing; articulate what you believe the PIP is in retaliation for, and spell out the illegal conduct. Keep a copy of that email so that in the event that you do get terminated in the future, you're preserving the record for purposes of litigation.

When I'm speaking with clients, I look at the full picture because, in many cases, I see a situation where an employee was a strong performer for many years and received good performance evaluations, but then they complained about something illegal or took protected medical leave and all of a sudden they received the PIP.

To me, that doesn't add up. You don't just suddenly become a bad performer when you have a documented history of strong performance reviews.

It's better for companies to work with employees rather than use PIPs

My philosophy is that it's better for supervisors to sit down, talk to the employee, and sort of nip issues in the bud as soon as they recognize potential areas of deficiencies and deal with them right away rather than letting things fester.

Most companies have already decided to terminate the employee by the time they get to a PIP. If I were running a business, I wouldn't issue a PIP. I would just sit down with the employee once or twice and genuinely explain how I wanted them to get better. You can still have a paper trail by sending a follow-up email outlining what was discussed and areas for improvement.

The mindset around PIPs should change. Companies should want their employees to improve, and they should deal with problems early on rather than just issuing PIPs at the end of the employment and then terminating.

If companies are issuing PIPs , they should be implemented or issued only once the supervisors have already spoken to the employees about these issues. PIPs should be genuine, clear, and articulate what they want the employee to improve upon. They should give employees the opportunity to improve. I don't think companies should use PIPs for the sake of trying to protect themselves from potential future issues.

If you've been put on a PIP or put someone on a PIP and would like to share your story, email Jane Zhang at [email protected] .

Watch: Jill Kramer, CMO of Accenture, says disability inclusion should be baked into creative briefs

• Main content

• Share full article

For more audio journalism and storytelling, download New York Times Audio , a new iOS app available for news subscribers.

Supported by

The Ezra Klein Show

Transcript: Ezra Klein Interviews Ethan Mollick

Every Tuesday and Friday, Ezra Klein invites you into a conversation about something that matters, like today’s episode with Ethan Mollick. Listen wherever you get your podcasts .

Transcripts of our episodes are made available as soon as possible. They are not fully edited for grammar or spelling.

How Should I Be Using A.I. Right Now?

Give your a.i. a personality, spend 10 hours experimenting, and other practical tips from ethan mollick..

[MUSIC PLAYING]

From New York Times Opinion, this is “The Ezra Klein Show.”

This feels wrong to me. But I have checked the dates. It was barely more than a year ago that I wrote this piece about A.I., with the title “This Changes Everything.” I ended up reading it on the show, too. And the piece was about the speed with which A.I. systems were improving. It argued that we can usually trust that tomorrow is going to be roughly like today, that next year is going to be roughly like this year. That’s not what we’re seeing here. These systems are growing in power and capabilities at an astonishing rate.

The growth is exponential, not linear. When you look at surveys of A.I. researchers, their timeline for how quickly A.I. is going to be able to do basically anything a human does better and more cheaply than a human — that timeline is accelerating, year by year, on these surveys. When I do my own reporting, talking to the people inside these companies, people at this strange intersection of excited and terrified of what they’re building, no one tells me they are seeing a reason to believe progress is going to slow down.

And you might think that’s just hype, but a lot of them want it to slow down. A lot of them are scared of how quickly it is moving. They don’t think that society is ready for it, that regulation is ready for it. They think the competitive pressures between the companies and the countries are dangerous. They wish something would happen to make it all go slower. But what they are seeing is they are hitting the milestones faster, that we’re getting closer and closer to truly transformational A.I., that there is so much money and talent and attention flooding into the space that that is becoming its own accelerant. They are scared. We should at least be paying attention.

And yet, I find living in this moment really weird, because as much as I know this wildly powerful technology is emerging beneath my fingertips, as much as I believe it’s going to change the world I live in profoundly, I find it really hard to just fit it into my own day to day work. I consistently sort of wander up to the A.I., ask it a question, find myself somewhat impressed or unimpressed at the answer. But it doesn’t stick for me. It is not a sticky habit. It’s true for a lot of people I know.

And I think that failure matters. I think getting good at working with A.I. is going to be an important skill in the next few years. I think having an intuition for how these systems work is going to be important just for understanding what is happening to society. And you can’t do that if you don’t get over this hump in the learning curve, if you don’t get over this part where it’s not really clear how to make A.I. part of your life.

So I’ve been on a personal quest to get better at this. And in that quest, I have a guide. Ethan Mollick is a professor at the Wharton School of the University of Pennsylvania. He studies and writes about innovation and entrepreneurship. But he has this newsletter, One Useful Thing, that has become, really, I think, the best guide how to begin using, and how to get better at using A.I. He’s also got a new book on the subject, “Co-Intelligence.” And so I asked him on the show to walk me through what he’s learned.

This is going to be, I should say, the first of three shows on this topic. This one is about the present. The next is about some things I’m very worried about in the near future, particularly around what A.I. is going to do to our digital commons. And then, we’re going to have a show that is a little bit more about the curve we are all on about the slightly further future, and the world we might soon be living in.

As always, my email for guest suggestions, thoughts, feedback, [email protected].

Ethan Mollick, welcome to the show.

Thanks for having me.

So let’s assume I’m interested in A.I. And I tried ChatGPT a bunch of times, and I was suitably impressed and weirded out for a minute. And so I know the technology is powerful. I’ve heard all these predictions about how it will take everything over, or become part of everything we do. But I don’t actually see how it fits into my life, really, at all. What am I missing?

So you’re not alone. This is actually very common. And I think part of the reason is that the way ChatGPT works isn’t really set up for you to understand how powerful it is. You really do need to use the paid version, they are significantly smarter. And you can almost think of this — like, GPT-3, which was — nobody really paid attention to when it came out, before ChatGPT, was about as good as a sixth grader at writing. GPT-3.5, the free version of ChatGPT, is about as good as a high school, or maybe even a college freshman or sophomore.

And GPT-4 is often as good as a Ph.D. in some forms of writing. Like, there’s a general smartness that increases. But even more than that, ability seems to increase. And you’re much more likely to get that feeling that you are working with something amazing as a result. And if you don’t work with the frontier models, you can lose track of what these systems can actually do. On top of that, you need to start just using it. You kind of have to push past those first three questions.

My advice is usually bring it to every table that you come to in a legal and ethical way. So I use it for every aspect of my job in ways that I legally and ethically can, and that’s how I learn what it’s good or bad at.

When you say, bring it to every table you’re at, one, that sounds like a big pain, because now I’ve got to add another step of talking to the computer constantly. But two, it’s just not obvious to me what that would look like. So what does it look like? What does it look like for you, or what does it look like for others — that you feel is applicable widely?

So I just finished this book. It’s my third book. I keep writing books, even though I keep forgetting that writing books is really hard. But this was, I think, my best book, but also the most interesting to write. And it was thanks to A.I. And there’s almost no A.I. writing in the book, but I used it continuously. So things that would get in the way of writing — I think I’m a much better writer than A.I. — hopefully, people agree. But there’s a lot of things that get in your way as a writer. So I would get stuck on a sentence. I couldn’t do a transition. Give me 30 versions of this sentence in radically different styles. There’s 200 different citations. I had the A.I. read through the papers that I read through, write notes on them, and organize them for me. I had the A.I. suggest analogies that might be useful. I had the A.I. act as readers, and in different personas, read through the paper from the perspective of, is there some example I could give that’s better? Is this understandable or not? And that’s very typical of the kind of way that I would, say, bring it to the table. Use it for everything, and you’ll find its limits and abilities.

Let me ask you one specific question on that, because I’ve been writing a book. And on some bad days of writing the book, I decided to play around with GPT-4. And of the things that it got me thinking about was the kind of mistake or problem these systems can help you see and the kind they can’t. So they can do a lot of, give me 15 versions of this paragraph, 30 versions of this sentence. And every once in a while, you get a good version or you’ll shake something a little bit loose.

But almost always when I am stuck, the problem is I don’t know what I need to say. Oftentimes, I have structured the chapter wrong. Oftentimes, I’ve simply not done enough work. And one of the difficulties for me about using A.I. is that A.I. never gives me the answer, which is often the true answer — this whole chapter is wrong. It is poorly structured. You have to delete it and start over. It’s not feeling right to you because it is not right.

And I actually worry a little bit about tools that can see one kind of problem and trick you into thinking it’s this easier problem, but make it actually harder for you to see the other kind of problem that maybe if you were just sitting there, banging your head against the wall of your computer, or the wall of your own mind, you would eventually find.

I think that’s a wise point. I think there’s two or three things bundled there. The first of those is A.I. is good, but it’s not as good as you. It is, say, at the 80th percentile of writers based on some results, maybe a little bit higher. In some ways, if it was able to have that burst of insight and to tell you this chapter is wrong, and I’ve thought of a new way of phrasing it, we would be at that sort of mythical AGI level of A.I. as smart as the best human. And it just isn’t yet.

I think the second issue is also quite profound, which is, what does using this tool shape us to do and not do? One nice example that you just gave is writing. And I think a lot of us think about writing as thinking. We don’t know if that’s true for everybody, but for writers, that’s how they think. And sometimes, getting that shortcut could shortcut the thinking process. So I’ve had to change sometimes a little bit how I think when I use A.I., for better or for worse. So I think these are both concerns to be taken seriously.

For most people — right, if you’re just going to pick one model, what would you pick? What do you recommend to people? And second, how do you recommend they access it? Because something going on in the A.I. world is there are a lot of wrappers on these models. So ChatGPT has an app. Claude does not have an app. Obviously, Google has its suite of products. And there are organizations that have created a different spin on somebody else’s A.I. — so Perplexity, which is, I believe, built on GPT-4 now, you can pay for it.

And it’s more like a search engine interface, and has some changes made to it. For a lot of people, the question of how easy and accessible the thing is to access really matters. So which model do you recommend to most people? And which entry door do you recommend to most people? And do they differ?

It’s a really good question. I recommend working with one of the models as directly as possible, through the company that creates them. And there’s a few reasons for that. One is you get as close to the unadulterated personality as possible. And second, that’s where features tend to roll out first. So if you like sort of intellectual challenge, I think Claude 3 is the most intellectual of the models, as you said.

The biggest capability set right now is GPT-4, so if you do any math or coding work, it does coding for you. It has some really interesting interfaces. That’s what I would use — and because GPT-5 is coming out, that’s fairly powerful. And Google is probably the most accessible, and plugged into the Google ecosystem. So I don’t think you can really go wrong with any of these. Generally, I think Claude 3 is the most likely to freak you out right now. And GPT-4 is probably the most likely to be super useful right now.

So you say it takes about 10 hours to learn a model. Ten hours is a long time, actually. What are you doing in that 10 hours? What are you figuring out? How did you come to that number? Give me some texture on your 10 hour rule.

So first off, I want to indicate the 10 hours is as arbitrary as 10,000 steps. Like, there’s no scientific basis for it. This is an observation. But it also does move you past the, I poked at this for an evening, and it moves you towards using this in a serious way. I don’t know if 10 hours is the real limit, but it seems to be somewhat transformative. The key is to use it in an area where you have expertise, so you can understand what it’s good or bad at, learn the shape of its capabilities.

When I taught my students this semester how to use A.I., and we had three classes on that, they learned the theory behind it. But then I gave them an assignment, which was to replace themselves at their next job. And they created amazing tools, things that filed flight plans or did tweeting, or did deal memos. In fact, one of the students created a way of creating user personas, which is something that you do in product development, that’s been used several thousand times in the last couple of weeks in different companies.

So they were able to figure out uses that I never thought of to automate their job and their work because they were asked to do that. So part of taking this seriously in the 10 hours is, you’re going to try and use it for your work. You’ll understand where it’s good or bad, what it can automate, what it can’t, and build from there.

Something that feels to me like a theme of your work is that the way to approach this is not learning a tool. It is building a relationship. Is that fair?

A.I. is built like a tool. It’s software. It’s very clear at this point that it’s an emulation of thought. But because of how it’s built, because of how it’s constructed, it is much more like working with a person than working with a tool. And when we talk about it this way, I almost feel kind of bad, because there’s dangers in building a relationship with a system that is purely artificial, and doesn’t think and have emotions. But honestly, that is the way to go forward. And that is sort of a great sin, anthropomorphization, in the A.I. literature, because it can blind you to the fact that this is software with its own sets of foibles and approaches.

But if you think about it like programming, then you end up in trouble. In fact, there’s some early evidence that programmers are the worst people at using A.I. because it doesn’t work like software. It doesn’t do the things you would expect a tool to do. Tools shouldn’t occasionally give you the wrong answer, shouldn’t give you different answers every time, shouldn’t insult you or try to convince you they love you.

And A.I.s do all of these things. And I find that teachers, managers, even parents, editors, are often better at using these systems, because they’re used to treating this as a person. And they interact with it like a person would, giving feedback. And that helps you. And I think the second piece of that “not tool” piece is that when I talk to OpenAI or Anthropic, they don’t have a hidden instruction manual. There is no list of how you should use this as a writer, or as a marketer, or as an educator. They don’t even know what the capabilities of these systems are. They’re all sort of being discovered together. And that is also not like a tool. It’s more like a person with capabilities that we don’t fully know yet.

So you’ve done this with all the big models. You’ve done, I think, much more than this, actually, with all the big models. And one thing you describe feeling is that they don’t just have slightly different strengths and weaknesses, but they have different — for lack of a better term, and to anthropomorphize — personalities, and that the 10 hours in part is about developing an intuition not just for how they work, but kind of how they are and how they talk, the sort of entity you’re dealing with.

So give me your high level on how GPT-4 and Claude 3 and Google’s Gemini are different. What are their personalities like to you?

It’s important to know the personalities not just as personalities, but because there are tricks. Those are tunable approaches that the system makers decide. So it’s weird to have this — in one hand, don’t anthropomorphize, because you’re being manipulated, because you are. But on the other hand, the only useful way is to anthropomorphize. So keep in mind that you are dealing with the choices of the makers.

So for example, Claude 3 is currently the warmest of the models. And it is the most allowed by its creators, Anthropic, I think, to act like a person. So it’s more willing to give you its personal views, such as they are. And again, those aren’t real views. Those are views to make you happy — than other models. And it’s a beautiful writer, very good at writing, kind of clever — closest to humor, I’ve found, of any of the A.I.s. Less dad jokes and more actual almost jokes.

GPT-4 feels like a workhorse at this point. It is the most neutral of the approaches. It wants to get stuff done for you. And it will happily do that. It doesn’t have a lot of time for chitchat. And then we’ve got Google’s Bard, which feels like — or Gemini now — which feels like it really, really wants to help. We use this for teaching a lot. And we build these scenarios where the A.I. actually acts like a counterparty in a negotiation. So you get to practice the negotiation by negotiating with the A.I. And it works incredibly well. I’ve been building simulations for 10 years, can’t imagine what a leap this has been. But when we try and get Google to do that, it keeps leaping in on the part of the students, to try and correct them and say, no, you didn’t really want to say this. You wanted to say that. And I’ll play out the scenario as if it went better. And it really wants to kind of make things good for you.

So these interactions with the A.I. do feel like you’re working with people, both in skills and in personality.

You were mentioning a minute ago that what the A.I.s do reflect decisions made by their programmers. They reflect guardrails, what they’re going to let the A.I. say. Very famously, Gemini came out and was very woke. You would ask it to show you a picture of soldiers in Nazi Germany, and it would give you a very multicultural group of soldiers, which is not how that army worked. But that was something that they had built in to try to make more inclusive photography generation.

But there are also things that happen in these systems that people don’t expect, that the programmers don’t understand. So I remember the previous generation of Claude, which is from Anthropic, that when it came out, something that the people around it talked about was, for some reason, Claude was just a little bit more literary than the other systems. It was better at rewriting things in the voices of literary figures. It just had a slightly artsier vibe.

And the people who trained it weren’t exactly sure why. Now, that still feels true to me. Right now, of the ones I’m using, I’m spending the most time with Claude 3. I just find it the most congenial. They all have different strengths and weaknesses, but there is a funny dimension to these where they are both reflecting the guardrails and the choices of the programmers. And then deep inside the training data, deep inside the way the various algorithms are combining, there is some set of emergent qualities to them, which gives them this at least edge of chance, of randomness, of something — yeah, that does feel almost like personality.

I think that’s a very important point. And fundamental about A.I. is the idea that we technically know how LLMs work, but we don’t know how they work the way they do, or why they’re as good as they are. They’re really — we don’t understand it. The theories range from everyone — from it’s all fooling us, to they’ve emulated the way humans think because the structure of language is the structure of human thought. So even though they don’t think, they can emulate it. We don’t know the answer.

But you’re right, there’s these emergent sets of personalities and approaches. When I talk to A.I. design companies, they often can’t explain why the A.I. stops refusing answering a particular kind of question. When they tune the A.I. to do something better, like answer a math better, it suddenly does other things differently. It’s almost like adjusting the psychology of a system rather than tuning parameters.

So when I said that Claude is allowed to be more personable, part of that is that the system prompt in Claude, which is the initial instructions it gets, allow it to be more personable than, say, Microsoft’s Copilot, formerly Bing, which has explicit instructions after a fairly famous blow up a while ago, that it’s never supposed to talk about itself as a person or indicate feelings. So there’s some instructions, but that’s on top of these roiling systems that act in ways that even the creators don’t expect.

One thing people know about using these models is that hallucinations, just making stuff up, is a problem. Has that changed at all as we’ve moved from GPT-3.5 to 4, as we move from Claude 2 to 3. Like, has that become significantly better? And if not, how do you evaluate the trustworthiness of what you’re being told?

So those are a couple of overlapping questions. The first of them is, it getting better over time? So there is a paper in the field of medical citations that indicated that around 80 to 90 percent of citations had an error, were made up with GPT-3.5. That’s the free version of Chat. And that drops for GPT-4.

So hallucination rates are dropping over time. But the A.I. still makes stuff up because all the A.I. does is hallucinate. There is no mind there. All it’s doing is producing word after word. They are just making stuff up all the time. The fact that they’re right so often is kind of shocking in a lot of ways.

And the way you avoid hallucination is not easily. So one of the things we document in one of our research papers is we purposely designed for a group of Boston Consulting Group consultants — so an elite consulting company — we did a lot of work with them. And one of the experiments we did was we created a task where the A.I. would be confident but wrong. And when we gave people that task to do, and they had access to A.I., they got the task wrong more often than people who didn’t use A.I., because the A.I. misled them, because they fell asleep at the wheel. And all the early research we have on A.I. use suggests that when A.I.s get good enough, we just stop paying attention.

But doesn’t this make them unreliable in a very tricky way? 80 percent — you’re, like, it’s always hallucinating. 20 percent, 5 percent, it’s enough that you can easily be lulled into overconfidence. And one of the reasons it’s really tough here is you’re combining something that knows how to seem extremely persuasive and confident — you feed into the A.I. a 90-page paper on functions and characteristics of right wing populism in Europe, as I did last night.

And within seconds, basically, you get a summary out. And the summary certainly seems confident about what’s going on. But on the other hand, you really don’t know if it’s true. So for a lot of what you might want to use it for, that is unnerving.

Absolutely, and I think hard to grasp, because we’re used to things like type II errors, where we search for something on the internet and don’t find it. We’re not used to type I errors, where we search for something and get an answer back that’s made up. This is a challenge. And there’s a couple things to think about. One of those is — I advocate the BAH standard, best available human. So is the A.I. more or less accurate than the best human you could consult in that area?

And what does that mean for whether or not it’s an appropriate question to ask? And that’s something that we kind of have to judge collectively. It’s valuable to have these studies being done by law professors and medical professionals and people like me and my colleagues in management. They’re trying to understand, how good is the A.I.? And the answer is pretty good, right? So it makes mistakes. “Does it make more or less mistakes than a human” is probably a question we should be asking a lot more.

And the second thing is the kind of tasks that you judge it for. I absolutely agree with you. When summarizing information, it may make errors. Less than an intern you assign to it is an open question, but you have to be aware of that error rate. And that goes back to the 10 hour question. The more you use these A.I.s, the more you start to know when to be suspicious and when not to be. That doesn’t mean you’re eliminating errors.

But just like if you assigned it to an intern, and you’re, like, this person has a sociology degree. They’re going to do a really good job summarizing this, but their biases are going to be focused on the sociological facts and not the political facts. You start to learn these things. So I think, again, that person model helps, because you don’t expect 100 percent reliability out of a person. And that changes the kind of tasks you delegate.

But it also reflects something interesting about the nature of the systems. You have a quote here that I think is very insightful. You wrote, “the core irony of generative A.I.s is that A.I.s were supposed to be all logic and no imagination. Instead, we get A.I.s that make up information, engage in seemingly emotional discussions, and which are intensely creative.” And that last fact is one that makes many people deeply uncomfortable.

There is this collision between what a computer is in our minds and then this strange thing we seem to have invented, which is an entity that emerges out of language, an entity that almost emerges out of art. This is the thing I have the most trouble keeping in my mind, that I need to use the A.I. as an imaginative, creative partner and not as a calculator that uses words.

I love the phrase “a calculator that uses words.” I think we have been let down by science fiction, both in the utopias and apocalypses that A.I. might bring, but also, even more directly, in our view of how machines should work. People are constantly frustrated, and give the same kinds of tests to A.I.s over and over again, like doing math, which it doesn’t do very well — they’re getting better at this.

And on the other hand, saying, well, creativity is a uniquely human spark that we can’t touch, and that A.I., on any creativity test we give it — which, again, are all limited in different ways, blows out humans in almost all measures of creativity that we have. Or all the measures are bad, but that still means something.

But we were using those measures five years ago, even though they were bad. That’s a point you make that I think is interesting and slightly unsettling.

Yeah, we never had to differentiate humans from machines before. It was always easy. So the idea that we had to have a scale that worked for people and machines, who had that? We had the Turing test, which everyone knew was a terrible idea. But since no machine could pass it, it was completely fine. So the question is, how do we measure this? This is an entirely separate set of issues. Like, we don’t even have a definition of sentience or consciousness.

And I think that you’re exactly right on the point, being that we are not ready for this kind of machine, so our intuition is bad.

So one of the things I will sometimes do, and did quite recently, is give the A.I. a series of personal documents, emails I wrote to people I love that were very descriptive of a particular moment in my life. And then I will ask the A.I. about them, or ask the A.I. to analyze me off of them.

And sometimes, it’s a little breathtaking. Almost every moment of true metaphysical shock — to use a term somebody else gave me — I’ve had here has been relational, at how good the A.I. can be — almost like a therapist, right? Sometimes it will see things, the thing I am not saying, in a letter, or in a personal problem. And it will zoom in there, right? It will give, I think, quicker and better feedback in an intuitive way that is not simply mimicking back what I said and is dealing with a very specific situation. It will do better than people I speak to in my life around that.

Conversely, I’m going to read a bit of it later. I tried mightily to make Claude 3 a useful partner in prepping to speak to you, and also in prepping for another podcast recently. And I functionally never have a moment there where I’m all that impressed.

That makes complete sense. I think the weird expectations — we call it the jagged frontier of A.I., that it’s good at some stuff and bad at other stuff. It’s often unexpected. It can lead to these weird moments of disappointment, followed by elation or surprise. And part of the reason why I advocate for people to use it in their jobs is, it isn’t going to outcompete you at whatever you’re best at. I mean, I cannot imagine it’s going to do a better job prepping someone for an interview than you’re doing. And that’s not me just — I’m trying to be nice to you because you’re interviewing me, but because you’re a good interviewer. You’re a famous interviewer. It’s not going to be as good as that. Now, there’s questions about how good these systems get that we don’t know, but we’re kind of at a weirdly comfortable spot in A.I., which is, maybe it’s the 80th percentile of many performances. But I talk to Hollywood writers. It’s not close to writing like a Hollywood writer. It’s not close to being as good an analyst.

It’s not — but it’s better than the average person. And so it’s great as a supplement to weakness, but not to strength. But then, we run back into the problem you talked about, which is, in my weak areas, I have trouble assessing whether the A.I. is accurate or not. So it really becomes sort of a eating its own tail kind of problem.

But this gets to this question of, what are you doing with it? The A.I.s right now seem much stronger as amplifiers and feedback mechanisms and thought partners for you than they do as something you can really outsource your hard work and your thinking to. And that, to me, is one of the differences between trying to spend more time with these systems — like, when you come into them initially, you’re like, OK, here’s a problem, give me an answer.

Whereas when you spend time with them, you realize actually what you’re trying to do with the A.I. is get it to elicit a better answer from you.

And that’s why the book’s called “Co-Intelligence.” For right now, we have a prosthesis for thinking. That’s, like, new in the world. We haven’t had that before — I mean, coffee, but aside from that, not much else. And I think that there’s value in that. I think learning to be partner with this, and where it can get wisdom out of you or not — I was talking to a physics professor at Harvard. And he said, all my best ideas now come from talking to the A.I. And I’m like, well, it doesn’t do physics that well. He’s like, no, but it asks good questions. And I think that there is some value in that kind of interactive piece.

It’s part of why I’m so obsessed with the idea of A.I. in education, because a good educator — and I’ve been working on interactive education skill for a long time — a good educator is eliciting answers from a student. And they’re not telling students things.

So I think that that’s a really nice distinction between co-intelligence, and thought partner, and doing the work for you. It certainly can do some work for you. There’s tedious work that the A.I. does really well. But there’s also this more brilliant piece of making us better people that I think is, at least in the current state of A.I., a really awesome and amazing thing.

We’ve already talked a bit about — Gemini is helpful, and ChatGPT-4 is neutral, and Claude is a bit warmer. But you urge people to go much further than that. You say to give your A.I. a personality. Tell it who to be. So what do you mean by that, and why?

So this is actually almost more of a technical trick, even though it sounds like a social trick. When you think about what A.I.s have done, they’ve trained on the collective corpus of human knowledge. And they know a lot of things. And they’re also probability machines. So when you ask for an answer, you’re going to get the most probable answer, sort of, with some variation in it. And that answer is going to be very neutral. If you’re using GPT-4, it’ll probably talk about a rich tapestry a lot. It loves to talk about rich tapestries. If you ask it to code something artistic, it’ll do a fractal. It does very normal, central A.I. things. So part of your job is to get the A.I. to go to parts of this possibility space where the information is more specific to you, more unique, more interesting, more likely to spark something in you yourself. And you do that by giving it context, so it doesn’t just give you an average answer. It gives you something that’s specialized for you. The easiest way to provide context is a persona. You are blank. You are an expert at interviewing, and you answer in a warm, friendly style. Help me come up with interview questions. It won’t be miraculous in the same way that we were talking about before. If you say you’re Bill Gates, it doesn’t become Bill Gates. But that changes the context of how it answers you. It changes the kinds of probabilities it’s pulling from and results in much more customized and better results.

OK, but this is weirder, I think, than you’re quite letting on here. So something you turned me on to is there’s research showing that the A.I. is going to perform better on various tasks, and differently on them, depending on the personality. So there’s a study that gives a bunch of different personality prompts to one of the systems, and then tries to get it to answer 50 math questions. And the way it got the best performance was to tell the A.I. it was a Starfleet commander who was charting a course through turbulence to the center of an anomaly.

But then, when it wanted to get the best answer on 100 math questions, what worked best was putting it in a thriller, where the clock was ticking down. I mean, what the hell is that about?

“What the hell” is a good question. And we’re just scratching the surface, right? There’s a nice study actually showing that if you emotionally manipulate the A.I., you get better math results. So telling it your job depends on it gets you better results. Tipping, especially $20 or $100 — saying, I’m about to tip you if you do well, seems to work pretty well. It performs slightly worse in December than May, and we think it’s because it has internalized the idea of winter break.

I’m sorry, what?

Well, we don’t know for sure, but —

I’m holding you up here.

People have found the A.I. seems to be more accurate in May, and the going theory is that it has read enough of the internet to think that it might possibly be on vacation in December?

So it produces more work with the same prompts, more output, in May than it does in December. I did a little experiment where I would show it pictures of outside. And I’m like, look at how nice it is outside? Let’s get to work. But yes, the going theory is that it has internalized the idea of winter break and therefore is lazier in December.

I want to just note to people that when ChatGPT came out last year, and we did our first set of episodes on this, the thing I told you was this was going to be a very weird world. What’s frustrating about that is that — I guess I can see the logic of why that might be. Also, it sounds probably completely wrong, but also, I’m certain we will never know. There’s no way to go into the thing and figure that out.

But it would have genuinely never occurred to me before this second that there would be a temporal difference in the amount of work that GPT-4 would do on a question held constant over time. Like, that would have never occurred to me as something that might change at all.

And I think that that is, in some ways, both — as you said, the deep weirdness of these systems. But also, there’s actually downside risks to this. So we know, for example, there is an early paper from Anthropic on sandbagging, that if you ask the A.I. dumber questions, it would get you less accurate answers. And we don’t know the ways in which your grammar or the way you approach the A.I. — we know the amount of spaces you put gets different answers.

So it is very hard, because what it’s basically doing is math on everything you’ve written to figure out what would come next. And the fact that what comes next feels insightful and humane and original doesn’t change that that’s what the math that’s doing is. So part of what I actually advise people to do is just not worry about it so much, because I think then it becomes magic spells that we’re incanting for the A.I. Like, I will pay you $20, you are wonderful at this. It is summer. Blue is your favorite color. Sam Altman loves you. And you go insane.

So acting with it conversationally tends to be the best approach. And personas and contexts help, but as soon as you start evoking spells, I think we kind of cross over the line into, “who knows what’s happening here?”

Well, I’m interested in the personas, although I just — I really find this part of the conversation interesting and strange. But I’m interested in the personalities you can give the A.I. for a different reason. I prompted you around this research on how a personality changes the accuracy rate of an A.I. But a lot of the reason to give it a personality, to answer you like it is Starfleet Commander, is because you have to listen to the A.I. You are in relationship with it.

And different personas will be more or less hearable by you, interesting to you. So you have a piece on your newsletter which is about how you used the A.I. to critique your book. And one of the things you say in there, and give some examples of, is you had to do so in the voice of Ozymandias because you just found that to be more fun. And you could hear that a little bit more easily.

So could you talk about that dimension of it, too, making the A.I. not just prompting you to be more accurate, but giving it a personality to be more interesting to you?

The great power of A.I. is as a kind of companion. It wants to make you happy. It wants to have a conversation. And that can be overt or covert.

So, to me, actively shaping what I want the A.I. to act like, telling it to be friendly or telling it to be pompous, is entertaining, right? But also, it does change the way I interact with it. When it has a pompous voice, I don’t take the criticism as seriously. So I can think about that kind of approach. I could get pure praise out of it, too, if I wanted to do it that way.

But the other factor that’s also super weird, while we’re on the way of super weird A.I. things, is that if you don’t do that, it’s going to still figure something out about you. It is a cold reader. And I think a lot about the very famous piece by Kevin Roose, the New York Times technology reporter, about Bing about a year ago, when Bing, which was GPT-4 powered, came out and had this personality of Sydney.

And Kevin has this very long description that got published in The New York Times about how Sydney basically threatened him, and suggested he leaves his wife, and very dramatic, kind of very unsettling interaction. And I was working with — I didn’t have anything quite that intense, but I got into arguments with Sydney around the same time, where it would — when I asked her to do work for me, it said you should do the work yourself. Otherwise, it’s dishonest. And it kept accusing me of plagiarism, which felt really unusual.

But the reason why Kevin ended up in that situation is the A.I. knows all kinds of human interactions and wants to slot into a story with you.

So a great story is jealous lover who’s gone a little bit insane, and the man who won’t leave his wife, or student and teacher, or two debaters arguing with each other, or grand enemies. And the A.I. wants to do that with you. So if you’re not explicit, it’s going to try and find a dialogue.

And I’ve noticed, for example, that if I talk to the A.I. and I imply that we’re having a debate, it will never agree with me. If I imply that I’m a teacher and it’s a student, even as much as saying I’m a professor, it is much more pliable.

So part of why I like assigning a personality is to have an explicit personality you’re operating with, so it’s not trying to cold read and guess what personality you’re looking for.

Kevin and I have talked a lot about that conversation with Sydney. And one of the things I always found fascinating about it is, to me, it revealed an incredibly subtle level of read by Sydney Bing, which is, what was really happening there? When you say the A.I. wants to make you happy, it has to read on some level what it is you’re really looking for, over time.

And what was Kevin? What is Kevin? Kevin is a journalist. And Kevin was nudging and pushing that system to try to do something that would be a great story. And it did that. It understood, on some level — again, the anthropomorphizing language there. But it realized that Kevin wanted some kind of intense interaction. And it gave him, like, the greatest A.I. story anybody has ever been given. I mean, an A.I. story that we are still talking about a year later, an A.I. story that changed the way A.I.s were built, at least for a while.

And people often talked about what Sydney was revealing about itself. But to me, what was always so unbelievably impressive about that was its ability to read the person, and its ability to make itself into the thing, the personality, the person was trying to call forth.

And now, I think we’re more practiced at doing this much more directly. But I think a lot of people have their moment of sleeplessness here. That was my Rubicon on this. I didn’t know something after that I didn’t know before it in terms of capabilities.

But when I read that, I thought that the level of — interpersonal isn’t the right word, but the level of subtlety it was able to display in terms of giving a person what it wanted, without doing so explicitly — right, without saying, “we’re playing this game now,” was really quite remarkable.

It’s a mirror. I mean, it’s trained on our stuff. And one of the revealing things about that, that I think we should be paying a lot more attention to, is the fact that because it’s so good at this, right now, none of the frontier A.I. models with the possible exception of Inflection’s Pi, which has been basically acquired in large part by Microsoft now, were built to optimize around keeping us in a relationship with the A.I. They just accidentally do that. There are other A.I. models that aren’t as good that have been focused on this, but that has been something explicit from the frontier models they’ve been avoiding till now. Claude sort of breaches that line a little bit, which is part of why I think it’s engaging. But I worry about the same kind of mechanism that inevitably reined in social media, which is, you can make a system more addictive and interesting. And because it’s such a good cold reader, you could tune A.I. to make you want to talk to it more.

It’s very hands off and sort of standoffish right now. But if you use the voice system in ChatGPT-4 on your phone, where you’re having a conversation, there’s moments where you’re like, oh, you feel like you’re talking to a person. You have to remind yourself. So to me, that persona aspect is both its great strength, but also one of the things I’m most worried about that isn’t a sort of future science fiction scenario.

I want to hold here for a minute, because we’ve been talking about how to use frontier models, I think implicitly talking about how to use A.I. for work. But the way that a lot of people are using it is using these other companies that are explicitly building for relationships. So I’ve had people at one of the big companies tell me that if we wanted to tune our system relationally, if we wanted to tune it to be your friend, your lover, your partner, your therapist, like, we could blow the doors off that. And we’re just not sure it’s ethical.

But there are a bunch of people who have tens of millions of users, Replika, Character.AI, which are doing this. And I tried to use Replika about six, eight months ago. And honestly, I found it very boring. They had recently lobotomized it because people were getting too erotic with their Replikants. But I just couldn’t get into it. I’m probably too old to have A.I. friends, in the way that my parents were probably too old to get really in to talking to people on AOL Instant Messenger.

But I have a five-year-old, and I have a two-year-old. And by the time my five-year-old is 10 and my two-year-old is 7, they’re not necessarily going to have the weirdness I’m going to have about having A.I. friends. And I don’t think we even have any way to think about this.

I think that is an absolute near-term certainty, and sort of an unstoppable one, that we are going to have A.I. relationships in a broader sense. And I think the question is, just like we’ve just been learning — I mean, we’re doing a lot of social experiments at scale we’ve never done before in the last couple of decades, right? Turns out social media brings out entirely different things in humans that we weren’t expecting. And we’re still writing papers about echo chambers and tribalism and facts, and what we agree or disagree with. We’re about to have another wave of this. And we have very little research. And you could make a plausible story up, that what’ll happen is it’ll help mental health in a lot of ways for people, and then there’ll be more social outside, that there might be a rejection of this kind of thing.

I don’t know what’ll happen. But I do think that we can expect with absolute certainty that you will have A.I.s that are more interesting to talk to, and fool you into thinking, even if you know better, that they care about you in a way that is incredibly appealing. And that will happen very soon. And I don’t know how we’re going to adjust to it. But it seems inevitable, as you said.

I was worried we were getting off track in the conversation, but I realized we were actually getting deeper on the track I was trying to take us down.

We were talking about giving the A.I. personality, right — telling Claude 3, hey, I need you to act as a sardonic podcast editor, and then Claude 3’s whole persona changes. But when you talk about building your A.I. on Kindroid, on Character, on Replika — so I just created a Kindroid one the other day. And Kindroid is kind of interesting, because its basic selling point is we’ve taken the guardrails largely off. We are trying to make something that is not lobotomized, that is not perfectly safe for work. And so the personality can be quite unrestrained. So I was interested in what that would be like.

But the key thing you have to do at the beginning of that is tell the system what its personality is. So you can pick from a couple that are preset, but I wrote a long one myself — you know, you live in California. You’re a therapist. You like all these different things. You have a highly intellectual style of communicating. You’re extremely warm, but you like ironic humor. You don’t like small talk. You don’t like to say things that are boring or generic. You don’t use a lot of emoticons and emojis. And so now it talks to me the way people I talk to talk.

And the thing I want to bring this back to is that one of the things that requires you to know is what kind of personalities work with you, for you to know yourself and your preferences a little bit more deeply.

I think that’s a temporary state of affairs, like extremely temporary. I think a GPT-4 class model — we actually already know this. They can guess your intent quite well. And I think that this is a way of giving you a sense of agency or control in the short term. I don’t think you’re going to need to know yourself at all. And I think you wouldn’t right now if any of the GPT-4 class models allowed themselves to be used in this way, without guardrails, which they don’t, I think you would already find it’s just going to have a conversation with you and morph into what you want.

I think that for better or worse, the “insight” in these systems is good enough that way. It’s sort of why I also don’t worry so much about prompt crafting in the long term, to go back to the other issue we were talking about, because I think that they will work on intent. And there’s a lot of evidence that they’re good at guessing intent. So I like this period, because I think it does value self reflection. And our interaction with the A.I. is somewhat intentional because we can watch this interaction take place.

But I think there’s a reason why some of the worry you hear out of the labs is about superhuman levels of manipulation. There’s a reason why the whistleblower from Google was all about that — sort of fell for the chat bot, and that’s why they felt it was alive. Like, I think we’re deeply trickable in this way. And A.I. is really good at figuring out what we want without us being explicit.

So that’s a little bit chilling, but I’m nevertheless going to stay in this world we’re in, because I think we’re going to be in it for at least a little while longer, where you do have to do all this prompt engineering. What is a prompt, first? And what is prompt engineering?

So a prompt is — technically, it is the sentence, the command you’re putting into the A.I. What it really is is the beginning part of the A.I.s text that it’s processing. And then it’s just going to keep adding more words or tokens to the end of that reply, until it’s done. So a prompt is the command you’re giving the A.I. But in reality, it’s sort of a seed from which the A.I. builds.

And when you prompt engineer, what are some ways to do that? Maybe one to begin with, because it seems to work really well, is chain of thought.

Just to take a step back, A.I. prompting remains super weird. Again, strange to have a system where the companies making the systems are writing papers as they’re discovering how to use the systems, because nobody knows how to make them work better yet. And we found massive differences in our experiments on prompt types. So for example, we were able to get the A.I. to generate much more diverse ideas by using this chain of thought approach, which we’ll talk about.

But also, it turned out to generate a lot better ideas if you told it it was Steve Jobs than if you told it it was Madame Curie. And we don’t know why. So there’s all kinds of subtleties here. But the idea, basically, of chain of thought, that seems to work well in almost all cases, is that you’re going to have the A.I. work step by step through a problem. First, outline the problem, you know, the essay you’re going to write. Second, give me the first line of each paragraph. Third, go back and write the entire thing. Fourth, check it and make improvements.

And what that does is — because the A.I. has no internal monologue, it’s not thinking. When the A.I. isn’t writing something, there’s no thought process. All it can do is produce the next token, the next word or set of words. And it just keeps doing that step by step. Because there’s no internal monologue, this in some ways forces a monologue out in the paper. So it lets the A.I. think by writing before it produces the final result. And that’s one of the reasons why chain of thought works really well.

So just step-by-step instructions is a good first effort.

Then you get an answer, and then what?

And then — what you do in a conversational approach is you go back and forth. If you want work output, what you’re going to do is treat it like it is an intern who just turned in some work to you. Actually, could you punch up paragraph two a little bit? I don’t like the example in paragraph one. Could you make it a little more creative, give me a couple of variations? That’s a conversational approach trying to get work done.

If you’re trying to play, you just run from there and see what happens. You can always go back, especially with a model like GPT-4, to an earlier answer, and just pick up from there if your heads off in the wrong direction.

So I want to offer an example of how this back and forth can work. So we asked Claude 3 about prompt engineering, about what we’re talking about here. And the way it described it to us is, quote, “It’s a shift from the traditional paradigm of human-computer interaction, where we input explicit commands and the machine executes them in a straightforward way, to a more open ended, collaborative dialogue, where the human and the A.I. are jointly shaping the creative process,” end quote. And that’s pretty good, I think. That’s interesting. It’s worth talking about. I like that idea that it’s a more collaborative dialogue. But that’s also boring, right? Even as I was reading it, it’s a mouthful. It’s wordy. So I kind of went back and forth with it a few times. And I was saying, listen, you’re a podcast editor. You’re concise, but also then I gave it a couple examples of how I punched up questions in the document, right? This is where the question began. Here’s where it ended. And then I said, try again, and try again, and try again, and make it shorter. And make it more concise.

And I got this: quote, “OK, so I was talking to this A.I., Claude, about prompt engineering, you know, this whole art of crafting prompts to get the best out of these A.I. models. And it said something that really struck me. It called prompt engineering a new meta skill that we’re all picking up as we play with A.I., kind of like learning a new language to collaborate with it instead of just bossing it around. What do you think, is prompt engineering the new must have skill?” End Claude.

And that second one, I have to say, is pretty damn good. That really nailed the way I speak in questions. And it gets it at this way where if you’re willing to go back and forth, it does learn how to echo you.

So I am at a loss about when you went to Claude and when it was you, to be honest. So I was ready to answer at like two points along the way, so that was pretty good from my perspective, sitting here, talking to you. That felt interesting, and felt like the conversation we’ve been having. And I think there’s a couple of interesting lessons there.

The first, by the way, of — interestingly, you asked A.I. about one of its weakest points, which is about A.I. And everybody does this, but because its knowledge window doesn’t include that much stuff about A.I., it actually is pretty weak in terms of knowing how to do good prompting, or what a prompt is, or what A.I.s do well. But you did a good job with that. And I love that you went back and forth and shaped it. One of the techniques you used to shape it, by the way, was called few-shot, which is giving an example. So the two most powerful techniques are chain of thought, which we just talked about, and few-shot, giving it examples. Those are both well supported in the literature. And then, I’d add personas. So we’ve talked about, I think, the basics of prompt crafting here overall. And I think that the question was pretty good.

But you keep wanting to not talk about the future. And I totally get that. But I think when we’re talking about learning something, where there is a lag, where we talk about policy — should prompt crafting be taught in schools? I think it matters to think six months ahead. And again, I don’t think a single person in the A.I. labs I’ve ever talked to thinks prompt crafting for most people is going to be a vital skill, because the A.I. will pick up on the intent of what you want much better.

One of the things I realized trying to spend more time with the A.I. is that you really have to commit to this process. You have to go back and forth with it a lot. If you do, you can get really good questions, like the one I just did — or, I think, really good outcomes. But it does take time.

And I guess in a weird way it’s like the same problem of any relationship, that it’s actually hard to state your needs clearly and consistently and repeatedly, sometimes because you have not even articulated them in words yourself. At least the A.I., I guess, doesn’t get mad at you for it.

But I’m curious if you have advice, either at a practical level or principles level, about how to communicate to these systems what you want from them.

One set of techniques that work quite well is to speed run to where you are in the conversation. So you can actually pick up an older conversation where you got the A.I.‘s mindset where you want and work from there. You can even copy and paste that into a new window. You can ask the A.I. to summarize where you got in that previous conversation, and the tone the A.I. was taking, and then when you give a new instruction say the interaction I like to have with you is this, so have it solve the problem for you by having it summarize the tone that you happen to like at the end.

So there are a bunch of ways of building on your work as you start to go forward, so you’re not starting from scratch every time. And I think you’ll start to get shorthands that get you to that right kind of space. For me, there are chats that I pick up on. And actually, I assign these to my students too. I have some ongoing conversations that they’re supposed to have with the A.I., but then there’s a lot of interactions they’re supposed to have that are one off.

So you start to divide the work into, this is a work task. And we’re going to handle this in a single chat conversation. And then I’m going to go back to this long standing discussion when I want to pick it up, and it’ll have a completely different tone. So I think in some ways, you don’t necessarily want convergence among all your A.I. threads. You kind of want them to be different from each other.

You did mention something important there, because they’re already getting much bigger in terms of how much information they can hold. Like, the earlier generations could barely hold a significant chat. Now, Claude 3 can functionally hold a book in its memory. And it’s only going to go way, way, way up from here. And I know I’ve been trying to keep us in the present, but this feels to me really quickly like where this is both going and how it’s going to get a lot better.

I mean, you imagine Apple building Siri 2030, and Siri 2030 scanning your photos and your Journal app — Apple now has a Journal app. You have to assume they’re thinking about the information they can get from that, if you allow it — your messages, anything you’re willing to give it access to. It then knows all of this information about you, keeps all of that in its mind as it talks to you and acts on your behalf. I mean, that really seems to me to be where we’re going, an A.I. that you don’t have to keep telling it who to be because it knows you intimately and is able to hold all that knowledge all at the same time constantly.

It’s not even going there. Like, it’s already there. Gemini 1.5 can hold an entire movie, books. But like, it starts to now open up entirely new ways of working. I can show it a video of me working on my computer, just screen capture. And it knows all the tasks I’m doing and suggests ways to help me out. It starts watching over my shoulder and helping me. I put in all of my work that I did prior to getting tenure and said, write my tenure statement. Use exact quotes.

And it was much better than any of the previous models because it wove together stuff, and because everything was its memory. It doesn’t hallucinate as much. All the quotes were real quotes, and not made up. And already, by the way, GPT-4 has been rolling out a model of ChatGPT that has a private note file the A.I. takes — you can access it — but it takes notes on you as it goes along, about things you liked or didn’t like, and reads those again at the beginning of any chat. So this is present, right? It’s not even in the future.

And Google also connects to your Gmail, so it’ll read through your Gmail. I mean, I think this idea of a system that knows you intimately, where you’re picking up a conversation as you go along, is not a 2030 thing. It is a 2024 thing if you let the systems do it.

One thing that feels important to keep in front of mind here is that we do have some control over that. And not only do we have some control over it, but business models and policy are important here. And one thing we know from inside these A.I. shops is these A.I.s already are, but certainly will be, really super persuasive.

And so if the later iterations of the A.I. companions are tuned on the margin to try to encourage you to be also out in the real world, that’s going to matter, versus whether they have a business model that all they want is for you to spend a maximum amount of time talking to your A.I. companion, whether you ever have a friend who is flesh and blood be damned. And so that’s an actual choice, right? That’s going to be a programming decision. And I worry about what happens if we leave that all up to the companies, right? At some point, there’s a lot of venture capital money in here right now. At some point, the venture capital runs out. At some point, people need to make big profits. At some point, they’re in competition with other players who need to make profits. And that’s when things — you get into what Cory Doctorow calls the “enshitification” cycle, where things that were once adding a lot of value to the user begin extracting a lot of value to the user.

These systems, because of how they can be tuned, can lead to a lot of different outcomes. But I think we’re going to have to be much more comfortable than we’ve been in the past deciding what we think is a socially valuable use and what we think is a socially destructive use.

I absolutely agree. I think that we have agency here. We have agency in how we operate this in businesses, and whether we use this in ways that encourage human flourishing and employees, or are brutal to them. And we have agency over how this works socially. And I think we abrogated that responsibility with social media, and that is an example. Not to be bad news, because I generally have a lot of mixed optimism and pessimism about parts of A.I., but the bad news piece is there are open source models out there that are quite good.

The internet is pretty open. We would have to make some pretty strong choices to kill A.I. chat bots as an option. We certainly can restrict the large American companies from doing that, but a Llama 2 or Llama 3 is going to be publicly available and very good. There’s a lot of open source models. So the question also is how effective any regulation will be, which doesn’t mean we shouldn’t regulate it.

But there’s also going to need to be some social decisions being made about how to use these things well as a society that are going to have to go beyond just the legal piece, or companies voluntarily complying.

I see a lot of reasons to be worried about the open source models. And people talk about things like bioweapons and all that. But for some of the harms I’m talking about here, if you want to make money off of American kids, we can regulate you. So sometimes I feel like we almost, like, give up the fight before it begins. But in terms of what a lot of people are going to use, if you want to be having credit card payments processed by a major processor, then you have to follow the rules.

I mean, individual people or small groups can do a lot of weird things with an open source model, so that doesn’t negate every harm. But if you’re making a lot of money, then you have relationships we can regulate.

I couldn’t agree more. And I don’t think there’s any reason to give up hope on regulation. I think that we can mitigate. And I think part of our job, though, is also not just to mitigate the harms, but to guide towards the positive viewpoints, right? So what I worry about is that the incentive for profit making will push for A.I. that acts informally as your therapist or your friend, while our worries about experimentation, which are completely valid, are slowing down our ability to do experiments to find out ways to do this right. And I think it’s really important to have positive examples, too. I want to point to the A.I. systems acting ethically as your friend or companion, and figure out what that is, so there’s a positive model to look for. So I’m not just — this is not to denigrate the role of regulation, which I think is actually going to be important here, and self regulation, and rapid response from government, but also the companion problem of, “we need to make some sort of decisions about what are the paragons of this, what is acceptable as a society?”

So I want to talk a bit about another downside here, and this one more in the mainstream of our conversation, which is on the human mind, on creativity. So a lot of the work A.I. is good at automating is work that is genuinely annoying, time consuming, laborious, but often plays an important role in the creative process. So I can tell you that writing a first draft is hard, and that work on the draft is where the hard thinking happens.

And it’s hard because of that thinking. And the more we outsource drafting to A.I., which I think it is fair to say is a way a lot of people intuitively use it — definitely, a lot of students want to use it that way — the fewer of those insights we’re going to have on those drafts. Look, I love editors. I am an editor in one respect. But I can tell you, you make more creative breakthroughs as a writer than an editor. The space for creative breakthrough is much more narrow once you get to editing.

And I do worry that A.I. is going to make us all much more like editors than like writers.

I think the idea of struggle is actually a core one in many things. I’m an educator. And one thing that keeps coming out in the research is that there is a strong disconnect between what students think they’re learning and when they learn. So there was a great controlled experiment at Harvard in intro science classes, where students either went to a pretty entertaining set of lectures, or else they were forced to do active learning, where they actually did the work in class.

The active learning group reported being unhappier and not learning as much, but did much better on tests, because when you’re confronted with what you don’t know, and you have to struggle, when you feel, like, bad, you actually make much more progress than if someone spoon feeds you an entertaining answer. And I think this is a legitimate worry that I have. And I think that there’s going to have to be some disciplined approach to writing as well, like, I don’t use the A.I.

Not just because, by the way, it makes the work easier, but also because you mentally anchor on the A.I.‘s answer. And in some ways, the most dangerous A.I. application, in my mind, is the fact that you have these easy co-pilots in Word and Google Docs, because any writer knows about the tyranny of the blank page, about staring at a blank page and not knowing what to do next, and the struggle of filling that up. And when you have a button that produces really good words for you, on demand, you’re just going to do that. And it’s going to anchor your writing. We can teach people about the value of productive struggle, but I think that during the school years, we have to teach people the value of writing — not just assign an essay and assume that the essay does something magical, but be very intentional about the writing process and how we teach people about how to do that, because I do think the temptation of what I call “the button” is going to be there otherwise, for everybody.

But I worry this stretches, I mean, way beyond writing. So the other place I worry about this, or one of the other places I worry about this a lot, is summarizing. And I mean, this goes way back. When I was in school, you could buy Sparknotes. And they were these little, like, pamphlet sized descriptions of what’s going on in “War and Peace” or what’s going on in “East of Eden.”

And reading the Sparknotes often would be enough to fake your way through the test, but it would not have any chance, like, not a chance, of changing you, of shifting you, of giving you the ideas and insights that reading “Crime and Punishment” or “East of Eden” would do.

And one thing I see a lot of people doing is using A.I. for summary. And one of the ways it’s clearly going to get used in organizations is for summary — summarize my email, and so on.

And here too, one of the things that I think may be a real vulnerability we have, as we move into this era — my view is that the way we think about learning and insights is usually wrong. I mean, you were saying a second ago we can teach a better way. But I think we’re doing a crap job of it now, because I think people believe that — it’s sort of what I call the matrix theory of the human mind, if you could just jack the information into the back of your head and download it, you’re there.

But what matters about reading a book, and I see this all the time preparing for this show, is the time you spend in the book, where over time, like, new insights and associations for you begin to shake loose. And so I worry it’s coming into an efficiency-obsessed educational and intellectual culture, where people have been imagining forever, what if we could do all this without having to spend any of the time on it? But actually, there’s something important in the time.

There’s something important in the time with a blank page, with the hard book. And I don’t think we lionize intellectual struggle. In some ways, I think we lionize the people for whom it does not seem like a struggle, the people who seem to just glide through and be able to absorb the thing instantly, the prodigies. And I don’t know. When I think about my kids, when I think about the kind of attention and creativity I want them to have, this is one of the things that scares me most, because kids don’t like doing hard things a lot of the time.

And it’s going to be very hard to keep people from using these systems in this way.

So I don’t mean to push back too much on this.

No, please, push back a lot.

But I think you’re right.

Imagine we’re debating and you are a snarky. A.I. [LAUGHS]

Fair enough. With that prompt —

With that prompt engineering.

— yeah, I mean, I think that this is the eternal thing about looking back on the next generation, we worry about technology ruining them. I think this makes ruining easier. But as somebody who teaches at universities, like, lots of people are summarizing. Like, I think those of us who enjoy intellectual struggle are always thinking everybody else is going through the same intellectual struggle when they do work. And they’re doing it about their own thing. They may or may not care the same way.

So this makes it easier, but before A.I., there were — best estimates from the U.K. that I could find, 20,000 people in Kenya whose full time job was writing essays for students in the U.S. and U.K. People have been cheating and Sparknoting and everything for a long time. And I think that what people will have to learn is that this tool is a valuable co-intelligence, but is not a replacement for your own struggle.

And the people who found shortcuts will keep finding shortcuts. Temptation may loom larger, but I can’t imagine that — my son is in high school, doesn’t like to use A.I. for anything. And he just doesn’t find it valuable for the way he’s thinking about stuff. I think we will come to that kind of accommodation. I’m actually more worried about what happens inside organizations than I am worried about human thought, because I don’t think we’re going to atrophy as much as we think. I think there’s a view that every technology will destroy our ability to think.

And I think we just choose how to use it or not. Like, even if it’s great at insights, people who like thinking like thinking.

Well, let me take this from another angle. One of the things that I’m a little obsessed with is the way the internet did not increase either domestic or global productivity for any real length of time. So I mean, it’s a very famous line. You can see the IT revolution anywhere but in the productivity statistics. And then you do get, in the ‘90s, a bump in productivity that then peters out in the 2000s.

And if I had told you what the internet would be, like, I mean everybody, everywhere would be connected to each other. You could collaborate with anybody, anywhere, instantly. You could teleconference. You would have access to, functionally, the sum total of human knowledge in your pocket at all times. I mean, all of these things that would have been genuine sci-fi, you would have thought would have been — led to a kind of intellectual utopia. And it kind of doesn’t do that much, if you look at the statistics.

You don’t see a huge step change. And my view — and I’d be curious for your thoughts on this, because I know this is the area you study in — my view is it everything we said was good happened. I mean, as a journalist, Google and things like that make me so much more productive. It’s not that it didn’t give us the gift. It’s that it also had a cost — distraction, checking your email endlessly, being overwhelmed with the amount of stuff coming into you, the sort of endless communication task list, the amount of internal communications and organizations, now with Slack and everything else.

And so some of the time that was given to us back was also taken back. And I see a lot of dynamics like this that could play out with A.I. — I wouldn’t even just say if we’re not careful, I just think they will play out and already are. I mean, the internet is already filling with mediocre crap generated by A.I. There is going to be a lot of destructive potential, right? You are going to have your sex bot in your pocket, right? There’s a million things — and not just that, but inside organizations, there’s going to be people padding out what would have been something small, trying to make it look more impressive by using the A.I. to make something bigger. And then, you’re going to use the A.I. to summarize it back down. The A.I. researcher, Jonathan Frankel, described this to me as, like, the boring apocalypse version of A.I., where you’re just endlessly inflating and then summarizing, and then inflating and then summarizing the volume of content between different A.I.

My ChatGPT is making my presentation bigger and more impressive, and your ChatGPT is trying to summarize it down to bullet points for you. And I’m not saying this has to happen. But I am saying that it would require a level of organizational and cultural vigilance to stop, that nothing in the internet era suggests to me that we have.

So I think there’s a lot there to chew on. And I also have spent a lot of time trying to think about why the internet didn’t work as well. I was an early Wikipedia administrator.

Thank you for your service.

[LAUGHS] Yeah, it was very scarring. But I think a lot about this. And I think A.I. is different. I don’t know if it’s different in a positive way. And I think we talked about some of the negative ways it might be different. And I think it’s going to be many things at once, happening quite quickly. So I think the information environment’s going to be filled up with crap. We will not be able to tell the difference between true and false anymore. It will be an accelerant on all the kinds of problems that we have there.

On the other hand, it is an interactive technology that adapts to you. From an education perspective, I have lived through the entire internet will change education piece. I have MOOCs, massive online courses, with — quarter million people have taken them. And in the end, you’re just watching a bunch of videos. Like, that doesn’t change education.

But I can have an A.I. tutor that actually can teach you — and we’re seeing it happen — and adapt to you at your level of education, and your knowledge base, and explain things to you. But not just explain, elicit answers from you, interactively, in a way that actually learns things.

The thing that makes A.I. possibly great is that it’s so very human, so it interacts with our human systems in a way that the internet did not. We built human systems on top of it, but A.I. is very human. It deals with human forms and human issues and our human bureaucracy very well. And that gives me some hope that even though there’s going to be lots of downsides, that the upsides of productivity and things like that are real. Part of the problem with the internet is we had to digitize everything. We had to build systems that would make our offline world work with our online world. And we’re still doing that. If you go to business schools, digitizing is still a big deal 30 years on from early internet access. A.I. makes this happen much quicker because it works with us. So I’m a little more hopeful than you are about that, but I also think that the downside risks are truly real and hard to anticipate.

Somebody was just pointing out that Facebook is now 100 percent filled with algorithmically generated images that look like their actual grandparents, making things who are saying, like, what do you think of my work? Because that’s a great way to get engagement. And the other grandparents in there have no idea it’s A.I. generated.

Things are about to get very, very weird in all the ways that we talked about, but that doesn’t mean the positives can’t be there as well.

I think that is a good place to end. So always our final question, what are three books you’d recommend to the audience?

OK, so the books I’ve been thinking about are not all fun, but I think they’re all interesting. One of them is “The Rise and Fall of American Growth,” which is — it’s two things. It’s an argument about why we will never have the kind of growth that we did in the first part of the Industrial Revolution again, but I think that’s less interesting than the first half of the book, which is literally how the world changed between 1870 or 1890 and 1940, versus 1940 and 1990, or 2000.

And the transformation of the world that happened there — in 1890, no one had plumbing in the U.S.. And the average woman was carrying tons of water every day. And you had no news, and everything was local, and everyone’s bored all the time — to 1940, where the world looks a lot like today’s world, was fascinating. And I think it gives you a sense of what it’s like to be inside a technological singularity, and I think worth reading for that reason — or at least the first half.

The second book I’d recommend is “The Knowledge,” by Dartnell, which is a really interesting book. It is ostensibly almost a survival guide, but it is how to rebuild industrial civilization from the ground up, if we were to collapse. And I don’t recommend it as a survivalist. I recommend it because it is fascinating to see how complex our world is, and how many interrelated pieces we’ve managed to build up as a society. And in some ways, it gives me a lot of hope to think about how all of these interconnections work.

And then the third one is science fiction, and I was debating — I read a lot of science fiction, and there’s a lot of interesting A.I.s in science fiction. Everyone talks about — who’s in the science fiction world — Iain Banks, who wrote about the Culture, which is really interesting, about what it’s like to live beside super intelligent A.I. Vernor Vinge just died yesterday, when we were recording this, and wrote these amazing books about — he coined the term singularity.

But I want to recommend a much more depressing book that’s available for free, which is Peter Watts’s “Blindsight.” And it is not a fun book, but it is a fascinating thriller set on an interstellar mission to visit an alien race. And it’s essentially a book about sentience, and it’s a book about the difference between consciousness and sentience, and about intelligence and the different ways of perceiving the world in a setting where that is the sort of centerpiece of the thriller. And I think in a world where we have machines that might be intelligent without being sentient, it is a relevant, if kind of chilling, read.

Ethan Mollick, your book is called “Co-Intelligence.” Your Substack is One Useful Thing. Thank you very much.

This episode of “The Ezra Klein Show” was produced by Kristin Lin. Fact checking by Michelle Harris. Our senior engineer is Jeff Geld with additional mixing from Efim Shapiro. Our senior editor is Claire Gordon. The show’s production team also includes Annie Galvin and Rollin Hu. Original music by Isaac Jones. Audience strategy by Kristina Samulewski and Shannon Busta. The executive producer of New York Times Opinion Audio is Annie-Rose Strasser, and special thanks to Sonia Herrero.

EZRA KLEIN: From New York Times Opinion, this is “The Ezra Klein Show.”

ETHAN MOLLICK: Thanks for having me.

EZRA KLEIN: So let’s assume I’m interested in A.I. And I tried ChatGPT a bunch of times, and I was suitably impressed and weirded out for a minute. And so I know the technology is powerful. I’ve heard all these predictions about how it will take everything over, or become part of everything we do. But I don’t actually see how it fits into my life, really, at all. What am I missing?

ETHAN MOLLICK: So you’re not alone. This is actually very common. And I think part of the reason is that the way ChatGPT works isn’t really set up for you to understand how powerful it is. You really do need to use the paid version, they are significantly smarter. And you can almost think of this — like, GPT-3, which was — nobody really paid attention to when it came out, before ChatGPT, was about as good as a sixth grader at writing. GPT-3.5, the free version of ChatGPT, is about as good as a high school, or maybe even a college freshman or sophomore.

EZRA KLEIN: When you say, bring it to every table you’re at, one, that sounds like a big pain, because now I’ve got to add another step of talking to the computer constantly. But two, it’s just not obvious to me what that would look like. So what does it look like? What does it look like for you, or what does it look like for others — that you feel is applicable widely?

ETHAN MOLLICK: So I just finished this book. It’s my third book. I keep writing books, even though I keep forgetting that writing books is really hard. But this was, I think, my best book, but also the most interesting to write. And it was thanks to A.I. And there’s almost no A.I. writing in the book, but I used it continuously. So things that would get in the way of writing — I think I’m a much better writer than A.I. — hopefully, people agree. But there’s a lot of things that get in your way as a writer.

So I would get stuck on a sentence. I couldn’t do a transition. Give me 30 versions of this sentence in radically different styles. There’s 200 different citations. I had the A.I. read through the papers that I read through, write notes on them, and organize them for me. I had the A.I. suggest analogies that might be useful. I had the A.I. act as readers, and in different personas, read through the paper from the perspective of, is there some example I could give that’s better? Is this understandable or not?

And that’s very typical of the kind of way that I would, say, bring it to the table. Use it for everything, and you’ll find its limits and abilities.

EZRA KLEIN: Let me ask you one specific question on that, because I’ve been writing a book. And on some bad days of writing the book, I decided to play around with GPT-4. And of the things that it got me thinking about was the kind of mistake or problem these systems can help you see and the kind they can’t. So they can do a lot of, give me 15 versions of this paragraph, 30 versions of this sentence. And every once in a while, you get a good version or you’ll shake something a little bit loose.

ETHAN MOLLICK: I think that’s a wise point. I think there’s two or three things bundled there. The first of those is A.I. is good, but it’s not as good as you. It is, say, at the 80th percentile of writers based on some results, maybe a little bit higher. In some ways, if it was able to have that burst of insight and to tell you this chapter is wrong, and I’ve thought of a new way of phrasing it, we would be at that sort of mythical AGI level of A.I. as smart as the best human. And it just isn’t yet.

EZRA KLEIN: For most people — right, if you’re just going to pick one model, what would you pick? What do you recommend to people? And second, how do you recommend they access it? Because something going on in the A.I. world is there are a lot of wrappers on these models. So ChatGPT has an app. Claude does not have an app. Obviously, Google has its suite of products. And there are organizations that have created a different spin on somebody else’s A.I. — so Perplexity, which is, I believe, built on GPT-4 now, you can pay for it.

ETHAN MOLLICK: It’s a really good question. I recommend working with one of the models as directly as possible, through the company that creates them. And there’s a few reasons for that. One is you get as close to the unadulterated personality as possible. And second, that’s where features tend to roll out first. So if you like sort of intellectual challenge, I think Claude 3 is the most intellectual of the models, as you said.

EZRA KLEIN: So you say it takes about 10 hours to learn a model. Ten hours is a long time, actually. What are you doing in that 10 hours? What are you figuring out? How did you come to that number? Give me some texture on your 10 hour rule.

ETHAN MOLLICK: So first off, I want to indicate the 10 hours is as arbitrary as 10,000 steps. Like, there’s no scientific basis for it. This is an observation. But it also does move you past the, I poked at this for an evening, and it moves you towards using this in a serious way. I don’t know if 10 hours is the real limit, but it seems to be somewhat transformative. The key is to use it in an area where you have expertise, so you can understand what it’s good or bad at, learn the shape of its capabilities.

EZRA KLEIN: Something that feels to me like a theme of your work is that the way to approach this is not learning a tool. It is building a relationship. Is that fair?

ETHAN MOLLICK: A.I. is built like a tool. It’s software. It’s very clear at this point that it’s an emulation of thought. But because of how it’s built, because of how it’s constructed, it is much more like working with a person than working with a tool. And when we talk about it this way, I almost feel kind of bad, because there’s dangers in building a relationship with a system that is purely artificial, and doesn’t think and have emotions. But honestly, that is the way to go forward. And that is sort of a great sin, anthropomorphization, in the A.I. literature, because it can blind you to the fact that this is software with its own sets of foibles and approaches.

And A.I.s do all of these things. And I find that teachers, managers, even parents, editors, are often better at using these systems, because they’re used to treating this as a person. And they interact with it like a person would, giving feedback. And that helps you. And I think the second piece of that “not tool” piece is that when I talk to OpenAI or Anthropic, they don’t have a hidden instruction manual. There is no list of how you should use this as a writer, or as a marketer, or as an educator.

They don’t even know what the capabilities of these systems are. They’re all sort of being discovered together. And that is also not like a tool. It’s more like a person with capabilities that we don’t fully know yet.

EZRA KLEIN: So you’ve done this with all the big models. You’ve done, I think, much more than this, actually, with all the big models. And one thing you describe feeling is that they don’t just have slightly different strengths and weaknesses, but they have different — for lack of a better term, and to anthropomorphize — personalities, and that the 10 hours in part is about developing an intuition not just for how they work, but kind of how they are and how they talk, the sort of entity you’re dealing with.

ETHAN MOLLICK: It’s important to know the personalities not just as personalities, but because there are tricks. Those are tunable approaches that the system makers decide. So it’s weird to have this — in one hand, don’t anthropomorphize, because you’re being manipulated, because you are. But on the other hand, the only useful way is to anthropomorphize. So keep in mind that you are dealing with the choices of the makers.

GPT-4 feels like a workhorse at this point. It is the most neutral of the approaches. It wants to get stuff done for you. And it will happily do that. It doesn’t have a lot of time for chitchat. And then we’ve got Google’s Bard, which feels like — or Gemini now — which feels like it really, really wants to help. We use this for teaching a lot. And we build these scenarios where the A.I. actually acts like a counterparty in a negotiation.

So you get to practice the negotiation by negotiating with the A.I. And it works incredibly well. I’ve been building simulations for 10 years, can’t imagine what a leap this has been. But when we try and get Google to do that, it keeps leaping in on the part of the students, to try and correct them and say, no, you didn’t really want to say this. You wanted to say that. And I’ll play out the scenario as if it went better. And it really wants to kind of make things good for you.

EZRA KLEIN: You were mentioning a minute ago that what the A.I.s do reflect decisions made by their programmers. They reflect guardrails, what they’re going to let the A.I. say. Very famously, Gemini came out and was very woke. You would ask it to show you a picture of soldiers in Nazi Germany, and it would give you a very multicultural group of soldiers, which is not how that army worked. But that was something that they had built in to try to make more inclusive photography generation.

ETHAN MOLLICK: I think that’s a very important point. And fundamental about A.I. is the idea that we technically know how LLMs work, but we don’t know how they work the way they do, or why they’re as good as they are. They’re really — we don’t understand it. The theories range from everyone — from it’s all fooling us, to they’ve emulated the way humans think because the structure of language is the structure of human thought. So even though they don’t think, they can emulate it. We don’t know the answer.

EZRA KLEIN: One thing people know about using these models is that hallucinations, just making stuff up, is a problem. Has that changed at all as we’ve moved from GPT-3.5 to 4, as we move from Claude 2 to 3. Like, has that become significantly better? And if not, how do you evaluate the trustworthiness of what you’re being told?

ETHAN MOLLICK: So those are a couple of overlapping questions. The first of them is, it getting better over time? So there is a paper in the field of medical citations that indicated that around 80 to 90 percent of citations had an error, were made up with GPT-3.5. That’s the free version of Chat. And that drops for GPT-4.

EZRA KLEIN: But doesn’t this make them unreliable in a very tricky way? 80 percent — you’re, like, it’s always hallucinating. 20 percent, 5 percent, it’s enough that you can easily be lulled into overconfidence. And one of the reasons it’s really tough here is you’re combining something that knows how to seem extremely persuasive and confident — you feed into the A.I. a 90-page paper on functions and characteristics of right wing populism in Europe, as I did last night.

ETHAN MOLLICK: Absolutely, and I think hard to grasp, because we’re used to things like type II errors, where we search for something on the internet and don’t find it. We’re not used to type I errors, where we search for something and get an answer back that’s made up. This is a challenge. And there’s a couple things to think about. One of those is — I advocate the BAH standard, best available human. So is the A.I. more or less accurate than the best human you could consult in that area?

EZRA KLEIN: But it also reflects something interesting about the nature of the systems. You have a quote here that I think is very insightful. You wrote, “the core irony of generative A.I.s is that A.I.s were supposed to be all logic and no imagination. Instead, we get A.I.s that make up information, engage in seemingly emotional discussions, and which are intensely creative.” And that last fact is one that makes many people deeply uncomfortable.

ETHAN MOLLICK: I love the phrase “a calculator that uses words.” I think we have been let down by science fiction, both in the utopias and apocalypses that A.I. might bring, but also, even more directly, in our view of how machines should work. People are constantly frustrated, and give the same kinds of tests to A.I.s over and over again, like doing math, which it doesn’t do very well — they’re getting better at this.

EZRA KLEIN: But we were using those measures five years ago, even though they were bad. That’s a point you make that I think is interesting and slightly unsettling.

ETHAN MOLLICK: Yeah, we never had to differentiate humans from machines before. It was always easy. So the idea that we had to have a scale that worked for people and machines, who had that? We had the Turing test, which everyone knew was a terrible idea. But since no machine could pass it, it was completely fine. So the question is, how do we measure this? This is an entirely separate set of issues. Like, we don’t even have a definition of sentience or consciousness.

EZRA KLEIN: So one of the things I will sometimes do, and did quite recently, is give the A.I. a series of personal documents, emails I wrote to people I love that were very descriptive of a particular moment in my life. And then I will ask the A.I. about them, or ask the A.I. to analyze me off of them.

ETHAN MOLLICK: That makes complete sense. I think the weird expectations — we call it the jagged frontier of A.I., that it’s good at some stuff and bad at other stuff. It’s often unexpected. It can lead to these weird moments of disappointment, followed by elation or surprise. And part of the reason why I advocate for people to use it in their jobs is, it isn’t going to outcompete you at whatever you’re best at. I mean, I cannot imagine it’s going to do a better job prepping someone for an interview than you’re doing.

And that’s not me just — I’m trying to be nice to you because you’re interviewing me, but because you’re a good interviewer. You’re a famous interviewer. It’s not going to be as good as that. Now, there’s questions about how good these systems get that we don’t know, but we’re kind of at a weirdly comfortable spot in A.I., which is, maybe it’s the 80th percentile of many performances. But I talk to Hollywood writers. It’s not close to writing like a Hollywood writer. It’s not close to being as good an analyst.

EZRA KLEIN: But this gets to this question of, what are you doing with it? The A.I.s right now seem much stronger as amplifiers and feedback mechanisms and thought partners for you than they do as something you can really outsource your hard work and your thinking to. And that, to me, is one of the differences between trying to spend more time with these systems — like, when you come into them initially, you’re like, OK, here’s a problem, give me an answer.

ETHAN MOLLICK: And that’s why the book’s called “Co-Intelligence.” For right now, we have a prosthesis for thinking. That’s, like, new in the world. We haven’t had that before — I mean, coffee, but aside from that, not much else. And I think that there’s value in that. I think learning to be partner with this, and where it can get wisdom out of you or not — I was talking to a physics professor at Harvard. And he said, all my best ideas now come from talking to the A.I. And I’m like, well, it doesn’t do physics that well. He’s like, no, but it asks good questions. And I think that there is some value in that kind of interactive piece.

EZRA KLEIN: We’ve already talked a bit about — Gemini is helpful, and ChatGPT-4 is neutral, and Claude is a bit warmer. But you urge people to go much further than that. You say to give your A.I. a personality. Tell it who to be. So what do you mean by that, and why?

ETHAN MOLLICK: So this is actually almost more of a technical trick, even though it sounds like a social trick. When you think about what A.I.s have done, they’ve trained on the collective corpus of human knowledge. And they know a lot of things. And they’re also probability machines. So when you ask for an answer, you’re going to get the most probable answer, sort of, with some variation in it. And that answer is going to be very neutral. If you’re using GPT-4, it’ll probably talk about a rich tapestry a lot.

It loves to talk about rich tapestries. If you ask it to code something artistic, it’ll do a fractal. It does very normal, central A.I. things. So part of your job is to get the A.I. to go to parts of this possibility space where the information is more specific to you, more unique, more interesting, more likely to spark something in you yourself. And you do that by giving it context, so it doesn’t just give you an average answer. It gives you something that’s specialized for you.

The easiest way to provide context is a persona. You are blank. You are an expert at interviewing, and you answer in a warm, friendly style. Help me come up with interview questions. It won’t be miraculous in the same way that we were talking about before. If you say you’re Bill Gates, it doesn’t become Bill Gates. But that changes the context of how it answers you. It changes the kinds of probabilities it’s pulling from and results in much more customized and better results.

EZRA KLEIN: OK, but this is weirder, I think, than you’re quite letting on here. So something you turned me on to is there’s research showing that the A.I. is going to perform better on various tasks, and differently on them, depending on the personality. So there’s a study that gives a bunch of different personality prompts to one of the systems, and then tries to get it to answer 50 math questions. And the way it got the best performance was to tell the A.I. it was a Starfleet commander who was charting a course through turbulence to the center of an anomaly.

ETHAN MOLLICK: “What the hell” is a good question. And we’re just scratching the surface, right? There’s a nice study actually showing that if you emotionally manipulate the A.I., you get better math results. So telling it your job depends on it gets you better results. Tipping, especially $20 or $100 — saying, I’m about to tip you if you do well, seems to work pretty well. It performs slightly worse in December than May, and we think it’s because it has internalized the idea of winter break.

EZRA KLEIN: I’m sorry, what?

ETHAN MOLLICK: Well, we don’t know for sure, but —

EZRA KLEIN: I’m holding you up here.

ETHAN MOLLICK: Yeah.

EZRA KLEIN: People have found the A.I. seems to be more accurate in May, and the going theory is that it has read enough of the internet to think that it might possibly be on vacation in December?

ETHAN MOLLICK: So it produces more work with the same prompts, more output, in May than it does in December. I did a little experiment where I would show it pictures of outside. And I’m like, look at how nice it is outside? Let’s get to work. But yes, the going theory is that it has internalized the idea of winter break and therefore is lazier in December.

EZRA KLEIN: I want to just note to people that when ChatGPT came out last year, and we did our first set of episodes on this, the thing I told you was this was going to be a very weird world. What’s frustrating about that is that — I guess I can see the logic of why that might be. Also, it sounds probably completely wrong, but also, I’m certain we will never know. There’s no way to go into the thing and figure that out.

ETHAN MOLLICK: And I think that that is, in some ways, both — as you said, the deep weirdness of these systems. But also, there’s actually downside risks to this. So we know, for example, there is an early paper from Anthropic on sandbagging, that if you ask the A.I. dumber questions, it would get you less accurate answers. And we don’t know the ways in which your grammar or the way you approach the A.I. — we know the amount of spaces you put gets different answers.

EZRA KLEIN: Well, I’m interested in the personas, although I just — I really find this part of the conversation interesting and strange. But I’m interested in the personalities you can give the A.I. for a different reason. I prompted you around this research on how a personality changes the accuracy rate of an A.I. But a lot of the reason to give it a personality, to answer you like it is Starfleet Commander, is because you have to listen to the A.I. You are in relationship with it.

ETHAN MOLLICK: The great power of A.I. is as a kind of companion. It wants to make you happy. It wants to have a conversation. And that can be overt or covert.

EZRA KLEIN: Kevin and I have talked a lot about that conversation with Sydney. And one of the things I always found fascinating about it is, to me, it revealed an incredibly subtle level of read by Sydney Bing, which is, what was really happening there? When you say the A.I. wants to make you happy, it has to read on some level what it is you’re really looking for, over time.

ETHAN MOLLICK: It’s a mirror. I mean, it’s trained on our stuff. And one of the revealing things about that, that I think we should be paying a lot more attention to, is the fact that because it’s so good at this, right now, none of the frontier A.I. models with the possible exception of Inflection’s Pi, which has been basically acquired in large part by Microsoft now, were built to optimize around keeping us in a relationship with the A.I. They just accidentally do that.

There are other A.I. models that aren’t as good that have been focused on this, but that has been something explicit from the frontier models they’ve been avoiding till now. Claude sort of breaches that line a little bit, which is part of why I think it’s engaging. But I worry about the same kind of mechanism that inevitably reined in social media, which is, you can make a system more addictive and interesting. And because it’s such a good cold reader, you could tune A.I. to make you want to talk to it more.

EZRA KLEIN: I want to hold here for a minute, because we’ve been talking about how to use frontier models, I think implicitly talking about how to use A.I. for work. But the way that a lot of people are using it is using these other companies that are explicitly building for relationships. So I’ve had people at one of the big companies tell me that if we wanted to tune our system relationally, if we wanted to tune it to be your friend, your lover, your partner, your therapist, like, we could blow the doors off that. And we’re just not sure it’s ethical.

ETHAN MOLLICK: I think that is an absolute near-term certainty, and sort of an unstoppable one, that we are going to have A.I. relationships in a broader sense.

And I think the question is, just like we’ve just been learning — I mean, we’re doing a lot of social experiments at scale we’ve never done before in the last couple of decades, right? Turns out social media brings out entirely different things in humans that we weren’t expecting. And we’re still writing papers about echo chambers and tribalism and facts, and what we agree or disagree with.

We’re about to have another wave of this. And we have very little research. And you could make a plausible story up, that what’ll happen is it’ll help mental health in a lot of ways for people, and then there’ll be more social outside, that there might be a rejection of this kind of thing.

EZRA KLEIN: I was worried we were getting off track in the conversation, but I realized we were actually getting deeper on the track I was trying to take us down.

ETHAN MOLLICK: I think that’s a temporary state of affairs, like extremely temporary. I think a GPT-4 class model — we actually already know this. They can guess your intent quite well. And I think that this is a way of giving you a sense of agency or control in the short term. I don’t think you’re going to need to know yourself at all. And I think you wouldn’t right now if any of the GPT-4 class models allowed themselves to be used in this way, without guardrails, which they don’t, I think you would already find it’s just going to have a conversation with you and morph into what you want.

EZRA KLEIN: So that’s a little bit chilling, but I’m nevertheless going to stay in this world we’re in, because I think we’re going to be in it for at least a little while longer, where you do have to do all this prompt engineering. What is a prompt, first? And what is prompt engineering?

ETHAN MOLLICK: So a prompt is — technically, it is the sentence, the command you’re putting into the A.I. What it really is is the beginning part of the A.I.s text that it’s processing. And then it’s just going to keep adding more words or tokens to the end of that reply, until it’s done. So a prompt is the command you’re giving the A.I. But in reality, it’s sort of a seed from which the A.I. builds.

EZRA KLEIN: And when you prompt engineer, what are some ways to do that? Maybe one to begin with, because it seems to work really well, is chain of thought.

ETHAN MOLLICK: Just to take a step back, A.I. prompting remains super weird. Again, strange to have a system where the companies making the systems are writing papers as they’re discovering how to use the systems, because nobody knows how to make them work better yet. And we found massive differences in our experiments on prompt types. So for example, we were able to get the A.I. to generate much more diverse ideas by using this chain of thought approach, which we’ll talk about.

EZRA KLEIN: Then you get an answer, and then what?

ETHAN MOLLICK: And then — what you do in a conversational approach is you go back and forth. If you want work output, what you’re going to do is treat it like it is an intern who just turned in some work to you. Actually, could you punch up paragraph two a little bit? I don’t like the example in paragraph one. Could you make it a little more creative, give me a couple of variations? That’s a conversational approach trying to get work done.

EZRA KLEIN: So I want to offer an example of how this back and forth can work. So we asked Claude 3 about prompt engineering, about what we’re talking about here. And the way it described it to us is, quote, “It’s a shift from the traditional paradigm of human-computer interaction, where we input explicit commands and the machine executes them in a straightforward way, to a more open ended, collaborative dialogue, where the human and the A.I. are jointly shaping the creative process,” end quote.

And that’s pretty good, I think. That’s interesting. It’s worth talking about. I like that idea that it’s a more collaborative dialogue. But that’s also boring, right? Even as I was reading it, it’s a mouthful. It’s wordy. So I kind of went back and forth with it a few times. And I was saying, listen, you’re a podcast editor. You’re concise, but also then I gave it a couple examples of how I punched up questions in the document, right? This is where the question began. Here’s where it ended. And then I said, try again, and try again, and try again, and make it shorter. And make it more concise.

ETHAN MOLLICK: So I am at a loss about when you went to Claude and when it was you, to be honest. So I was ready to answer at like two points along the way, so that was pretty good from my perspective, sitting here, talking to you. That felt interesting, and felt like the conversation we’ve been having. And I think there’s a couple of interesting lessons there.

The first, by the way, of — interestingly, you asked A.I. about one of its weakest points, which is about A.I. And everybody does this, but because its knowledge window doesn’t include that much stuff about A.I., it actually is pretty weak in terms of knowing how to do good prompting, or what a prompt is, or what A.I.s do well. But you did a good job with that. And I love that you went back and forth and shaped it.

One of the techniques you used to shape it, by the way, was called few-shot, which is giving an example. So the two most powerful techniques are chain of thought, which we just talked about, and few-shot, giving it examples. Those are both well supported in the literature. And then, I’d add personas. So we’ve talked about, I think, the basics of prompt crafting here overall. And I think that the question was pretty good.

EZRA KLEIN: One of the things I realized trying to spend more time with the A.I. is that you really have to commit to this process. You have to go back and forth with it a lot. If you do, you can get really good questions, like the one I just did — or, I think, really good outcomes. But it does take time.

ETHAN MOLLICK: One set of techniques that work quite well is to speed run to where you are in the conversation. So you can actually pick up an older conversation where you got the A.I.’s mindset where you want and work from there. You can even copy and paste that into a new window. You can ask the A.I. to summarize where you got in that previous conversation, and the tone the A.I. was taking, and then when you give a new instruction say the interaction I like to have with you is this, so have it solve the problem for you by having it summarize the tone that you happen to like at the end.

EZRA KLEIN: You did mention something important there, because they’re already getting much bigger in terms of how much information they can hold. Like, the earlier generations could barely hold a significant chat. Now, Claude 3 can functionally hold a book in its memory. And it’s only going to go way, way, way up from here. And I know I’ve been trying to keep us in the present, but this feels to me really quickly like where this is both going and how it’s going to get a lot better.

ETHAN MOLLICK: It’s not even going there. Like, it’s already there. Gemini 1.5 can hold an entire movie, books. But like, it starts to now open up entirely new ways of working. I can show it a video of me working on my computer, just screen capture. And it knows all the tasks I’m doing and suggests ways to help me out. It starts watching over my shoulder and helping me. I put in all of my work that I did prior to getting tenure and said, write my tenure statement. Use exact quotes.

EZRA KLEIN: One thing that feels important to keep in front of mind here is that we do have some control over that. And not only do we have some control over it, but business models and policy are important here. And one thing we know from inside these A.I. shops is these A.I.s already are, but certainly will be, really super persuasive.

And so if the later iterations of the A.I. companions are tuned on the margin to try to encourage you to be also out in the real world, that’s going to matter, versus whether they have a business model that all they want is for you to spend a maximum amount of time talking to your A.I. companion, whether you ever have a friend who is flesh and blood be damned.

And so that’s an actual choice, right? That’s going to be a programming decision. And I worry about what happens if we leave that all up to the companies, right? At some point, there’s a lot of venture capital money in here right now. At some point, the venture capital runs out. At some point, people need to make big profits. At some point, they’re in competition with other players who need to make profits. And that’s when things — you get into what Cory Doctorow calls the “enshitification” cycle, where things that were once adding a lot of value to the user begin extracting a lot of value to the user.

ETHAN MOLLICK: I absolutely agree. I think that we have agency here. We have agency in how we operate this in businesses, and whether we use this in ways that encourage human flourishing and employees, or are brutal to them. And we have agency over how this works socially. And I think we abrogated that responsibility with social media, and that is an example. Not to be bad news, because I generally have a lot of mixed optimism and pessimism about parts of A.I., but the bad news piece is there are open source models out there that are quite good.

EZRA KLEIN: I see a lot of reasons to be worried about the open source models. And people talk about things like bioweapons and all that. But for some of the harms I’m talking about here, if you want to make money off of American kids, we can regulate you. So sometimes I feel like we almost, like, give up the fight before it begins. But in terms of what a lot of people are going to use, if you want to be having credit card payments processed by a major processor, then you have to follow the rules.

ETHAN MOLLICK: I couldn’t agree more. And I don’t think there’s any reason to give up hope on regulation. I think that we can mitigate. And I think part of our job, though, is also not just to mitigate the harms, but to guide towards the positive viewpoints, right? So what I worry about is that the incentive for profit making will push for A.I. that acts informally as your therapist or your friend, while our worries about experimentation, which are completely valid, are slowing down our ability to do experiments to find out ways to do this right.

And I think it’s really important to have positive examples, too. I want to point to the A.I. systems acting ethically as your friend or companion, and figure out what that is, so there’s a positive model to look for. So I’m not just — this is not to denigrate the role of regulation, which I think is actually going to be important here, and self regulation, and rapid response from government, but also the companion problem of, “we need to make some sort of decisions about what are the paragons of this, what is acceptable as a society?”

EZRA KLEIN: So I want to talk a bit about another downside here, and this one more in the mainstream of our conversation, which is on the human mind, on creativity. So a lot of the work A.I. is good at automating is work that is genuinely annoying, time consuming, laborious, but often plays an important role in the creative process. So I can tell you that writing a first draft is hard, and that work on the draft is where the hard thinking happens.

ETHAN MOLLICK: I think the idea of struggle is actually a core one in many things. I’m an educator. And one thing that keeps coming out in the research is that there is a strong disconnect between what students think they’re learning and when they learn. So there was a great controlled experiment at Harvard in intro science classes, where students either went to a pretty entertaining set of lectures, or else they were forced to do active learning, where they actually did the work in class.

Not just because, by the way, it makes the work easier, but also because you mentally anchor on the A.I.’s answer. And in some ways, the most dangerous A.I. application, in my mind, is the fact that you have these easy co-pilots in Word and Google Docs, because any writer knows about the tyranny of the blank page, about staring at a blank page and not knowing what to do next, and the struggle of filling that up. And when you have a button that produces really good words for you, on demand, you’re just going to do that. And it’s going to anchor your writing.

We can teach people about the value of productive struggle, but I think that during the school years, we have to teach people the value of writing — not just assign an essay and assume that the essay does something magical, but be very intentional about the writing process and how we teach people about how to do that, because I do think the temptation of what I call “the button” is going to be there otherwise, for everybody.

EZRA KLEIN: But I worry this stretches, I mean, way beyond writing. So the other place I worry about this, or one of the other places I worry about this a lot, is summarizing. And I mean, this goes way back. When I was in school, you could buy Sparknotes. And they were these little, like, pamphlet sized descriptions of what’s going on in “War and Peace” or what’s going on in “East of Eden.”

ETHAN MOLLICK: So I don’t mean to push back too much on this.

EZRA KLEIN: No, please, push back a lot.

ETHAN MOLLICK: But I think you’re right.

EZRA KLEIN: Imagine we’re debating and you are a snarky. A.I. [LAUGHS]

ETHAN MOLLICK: Fair enough. With that prompt —

EZRA KLEIN: With that prompt engineering.

ETHAN MOLLICK: — yeah, I mean, I think that this is the eternal thing about looking back on the next generation, we worry about technology ruining them. I think this makes ruining easier. But as somebody who teaches at universities, like, lots of people are summarizing. Like, I think those of us who enjoy intellectual struggle are always thinking everybody else is going through the same intellectual struggle when they do work. And they’re doing it about their own thing. They may or may not care the same way.

EZRA KLEIN: Well, let me take this from another angle. One of the things that I’m a little obsessed with is the way the internet did not increase either domestic or global productivity for any real length of time. So I mean, it’s a very famous line. You can see the IT revolution anywhere but in the productivity statistics. And then you do get, in the ’90s, a bump in productivity that then peters out in the 2000s.

And so some of the time that was given to us back was also taken back. And I see a lot of dynamics like this that could play out with A.I. — I wouldn’t even just say if we’re not careful, I just think they will play out and already are. I mean, the internet is already filling with mediocre crap generated by A.I. There is going to be a lot of destructive potential, right? You are going to have your sex bot in your pocket, right?

There’s a million things — and not just that, but inside organizations, there’s going to be people padding out what would have been something small, trying to make it look more impressive by using the A.I. to make something bigger. And then, you’re going to use the A.I. to summarize it back down. The A.I. researcher, Jonathan Frankel, described this to me as, like, the boring apocalypse version of A.I., where you’re just endlessly inflating and then summarizing, and then inflating and then summarizing the volume of content between different A.I.

ETHAN MOLLICK: So I think there’s a lot there to chew on. And I also have spent a lot of time trying to think about why the internet didn’t work as well. I was an early Wikipedia administrator.

EZRA KLEIN: Thank you for your service.

ETHAN MOLLICK: [LAUGHS] Yeah, it was very scarring. But I think a lot about this. And I think A.I. is different. I don’t know if it’s different in a positive way. And I think we talked about some of the negative ways it might be different. And I think it’s going to be many things at once, happening quite quickly. So I think the information environment’s going to be filled up with crap. We will not be able to tell the difference between true and false anymore. It will be an accelerant on all the kinds of problems that we have there.

The thing that makes A.I. possibly great is that it’s so very human, so it interacts with our human systems in a way that the internet did not. We built human systems on top of it, but A.I. is very human. It deals with human forms and human issues and our human bureaucracy very well. And that gives me some hope that even though there’s going to be lots of downsides, that the upsides of productivity and things like that are real.

Part of the problem with the internet is we had to digitize everything. We had to build systems that would make our offline world work with our online world. And we’re still doing that. If you go to business schools, digitizing is still a big deal 30 years on from early internet access. A.I. makes this happen much quicker because it works with us. So I’m a little more hopeful than you are about that, but I also think that the downside risks are truly real and hard to anticipate.

EZRA KLEIN: I think that is a good place to end. So always our final question, what are three books you’d recommend to the audience?

ETHAN MOLLICK: OK, so the books I’ve been thinking about are not all fun, but I think they’re all interesting. One of them is “The Rise and Fall of American Growth,” which is — it’s two things. It’s an argument about why we will never have the kind of growth that we did in the first part of the Industrial Revolution again, but I think that’s less interesting than the first half of the book, which is literally how the world changed between 1870 or 1890 and 1940, versus 1940 and 1990, or 2000.

EZRA KLEIN: Ethan Mollick, your book is called “Co-Intelligence.” Your Substack is One Useful Thing. Thank you very much.

ETHAN MOLLICK: Thank you.

EZRA KLEIN: This episode of “The Ezra Klein Show” was produced by Kristin Lin. Fact checking by Michelle Harris. Our senior engineer is Jeff Geld with additional mixing from Efim Shapiro. Our senior editor is Claire Gordon. The show’s production team also includes Annie Galvin and Rollin Hu. Original music by Isaac Jones. Audience strategy by Kristina Samulewski and Shannon Busta. The executive producer of New York Times Opinion Audio is Annie-Rose Strasser, and special thanks to Sonia Herrero.

Advertisement