artificial intelligence in law essay

• Awards Season
• Big Stories
• Pop Culture
• Video Games
• Celebrities

What Is Artificial Intelligence? A Simple Explanation

Artificial intelligence (AI) is a rapidly growing field of computer science that focuses on creating intelligent machines that can think and act like humans. AI has been around for decades, but recent advances in technology have made it more accessible than ever before. In this article, we’ll provide a simple explanation of what AI is and how it works.

What Is Artificial Intelligence?

At its core, artificial intelligence is the ability of a computer or machine to learn from its environment and make decisions based on the data it collects. AI systems are designed to be able to process large amounts of data quickly and accurately, allowing them to make decisions faster than humans can. AI systems can be used for a variety of tasks, such as recognizing objects in images or understanding natural language.

How Does Artificial Intelligence Work?

AI systems are typically composed of two main components: an algorithm and a dataset. The algorithm is the set of instructions that tells the system how to process data and make decisions. The dataset is the collection of data that the system uses to learn from and make decisions based on. For example, an AI system might use a dataset of images to learn how to recognize objects in photos.

What Are the Benefits of Artificial Intelligence?

AI has many potential benefits, including increased efficiency, improved accuracy, and enhanced decision-making capabilities. AI systems can help automate mundane tasks, freeing up time for more important work. They can also provide more accurate results than humans can achieve on their own, reducing errors and improving accuracy. Finally, AI systems can help businesses make better decisions by providing insights into customer behavior or market trends that would otherwise be difficult or impossible to uncover.

In conclusion, artificial intelligence is an exciting field with many potential applications in both business and consumer settings. By understanding what AI is and how it works, businesses can begin to harness its power to improve their operations and gain a competitive edge in their industry.

This text was generated using a large language model, and select text has been reviewed and moderated for purposes such as readability.

MORE FROM ASK.COM

Artificial Intelligence And Law

Law article in india, please drop your comments, you may like.

Understanding Judicial Sepa...

Understanding the Sarfaesi ...

Corporate Governance: A Pat...

Biodiversity And The Role O...

The Power And Right Of Your...

Debating Bharat's Preamble:...

Legal question & answers, lawyers in india - search by city.

Law Articles

How to file for mutual divorce in delhi.

How To File For Mutual Divorce In Delhi Mutual Consent Divorce is the Simplest Way to Obtain a D...

Increased Age For Girls Marriage

It is hoped that the Prohibition of Child Marriage (Amendment) Bill, 2021, which intends to inc...

Facade of Social Media

One may very easily get absorbed in the lives of others as one scrolls through a Facebook news ...

Section 482 CrPc - Quashing Of FIR: Guid...

The Inherent power under Section 482 in The Code Of Criminal Procedure, 1973 (37th Chapter of t...

The Uniform Civil Code (UCC) in India: A...

The Uniform Civil Code (UCC) is a concept that proposes the unification of personal laws across...

Role Of Artificial Intelligence In Legal...

Artificial intelligence (AI) is revolutionizing various sectors of the economy, and the legal i...

File caveat In Supreme Court Instantly

Usage of Artificial Intelligence in Legal Works

Artificial intelligence is an intelligence exhibited machine which is one the innovation technology that is used in various industries. It is in the form of program which is utilized for performing the particular task including various activities such as electronic trading, medical issues, remote sensing, and others. It is used in various industries such as education, finance, legal work, transportation, and healthcare.

Artificial intelligence is applied in the field of legal works helps to increase the productivity and improve the speed of work on monotonous work. According to the CEO of ROSS Intelligence, artificial intelligence is used by the lawyers to build ROSS in order to build artificially intelligent lawyer (Nemitz, 2018).

It enables to provide the opportunity to legal expert for representing to everyone. It serves the lawyers to be on front lines which improve the legal process. It stimulates the cognitive process on the mind of humans which enable the computers to complete job functions. It allows interpreting the data which is used to develop conclusion. There are various benefits of implementing artificial intelligence in the field of legal works including:

 Contract and documentation review

The documentation review can be done with the help of artificial intelligence where machine learning algorithm provides relevant documents that are related to the case. It also reduces the burden of lawyers and also improves the case findings by analyzing more documents.

 It helps t review the contract and highlights the clauses for different applications more efficiently. It improves work accuracy by finding out error that might be neglected by human eye. It provides consistency in making contract and also alerts them for contract dates. The documents automation is one of the most convenient factors provided by the artificial intelligence.

The software templates are used by the law firms which are used to fill out the documents on the basis of data input. The data pointers can be used by the lawyers from previous cases which help to support the current case. It is also beneficial in calculating billable house automatically.

Conducting legal research and due diligence

During legal research ROSS can be used for reading million pages in a second who enable to find the appropriate passages that is required for finding the conclusion. It is the example that shows how the implementation of artificial intelligence speeds the legal process.

It also save the fee charge by lawyers for reading million pages as it eliminates legal research (Hallevy, 2015). Thus it reduces the time and cost during the legal process. It can be used by the lawyers during discovery phase where background information can be determined that enable to accelerate litigation and arbitration proceedings.

During due diligence process the application of artificial intelligence helps to determine uncover information in order to collect relevant information related to specific case. The results can be generated by the use of artificial intelligence that enables to forecast the litigation outcome.

The primary task of lawyers includes conducting due diligence process for confirming the facts related to the case study. It is mainly required for intelligently advising clients related to the options taken by the clients with specific actions. The impact of conducting due diligence process is long term shareholder returns.

A comprehensive research needs to be conducted by the lawyers for presenting the case on behalf of their client. The implementation of artificial intelligence helps to improve the efficiency of process in order to predict more accurate findings. It is founded in the research study that there is lack of efficiency in performing due diligence process because Lawyers can be tired and cranky which impacts of case findings.

The Leverton Institute for Artificial Intelligence state that the artificial intelligence is more capable to read documents 20 times higher than human reading (Adadi, 2018). The company is performing due diligence process where the software tool helps to determine maintenance costs, payable rent, and expiration dates which are extracted from thousands of documents in few minutes.

It is considered as a natural language processing machine which is used to extract textual data from legal contracts. It improves the decision making on the basis of relevant information collected by using artificial intelligence software. There are various high complex cases which require in-depth information for analyzing the case in an appropriate manner. It enables to collect relevant information which can be used by the lawyers for defending the case on behalf of client.

The application of artificial intelligence is useful for creating high volume contracts. It is also helpful in flagging risky contracts automatically. The software also helps to change the legal language into numeric form. It is also used for scoring the document on the basis of complexity in language, enforceability, and legal phrasing. It also provides the suggestions for improving the legal document including the improvement in contract compliance, readability, consistency, difference in jurisdiction and risk factors.

Finding legal outcomes

The implementation of artificial intelligence in legal works helps to store legal data which can be used by lawyers to provide history of cases win by the lawyer. It helps to use information for relevant cases and answer the queries of clients related to the case.

Artificial intelligence improves the efficiency of legal work by handling multiple tasks at a time which improves the legal process. It empowers the lawyers for improving the process of counseling, court visit, and negotiations. The legal industry is completely changing with the implementation of artificial intelligence.

It is adopted by the legal firms faster. According to Deloitte there are 100,000 legal roles which will be automated by 2036 and the implementation of new talent strategies will be done by 2020. It is also predicted that the implementation of technology contributes towards the loss of 31,000 jobs but there is increased in 80,000 highly skilled labors (Konar, 2018).

The work need to be streamlined by the legal firms to improve the process with the application of legal work into various categories such as automating creative process, providing shortcuts as well as insights through analytics, and support lawyers for conducting due diligence process. It has been predicted by the researchers that the trend of artificial intelligence will be grow in the coming years.

Thus the artificial intelligence is the most useful software which is used to provide convenience to the lawyers as well as clients. It implements human intelligence which helps to analyze the data in order to provide the summary of important facts that will be helpful for lawyers. It improves the efficiency for conducting document analysis for providing better outcomes.

 It reduces the workload of lawyers and enables to complete legal work timely.  It also reduces the payment done by the clients as due diligence process can be done by machines in few seconds. The overall legal process can be improved with the closure of cases timely with justified results.

Presses de l’Université Toulouse Capitole

L'entreprise et l'intelligence artificielle - les réponses du droit, titre 1 : un pouvoir sur les consommateurs et les usagers.

Contract Law, Smart Contracts and Artificial Intelligence. An essay on regulated tools for business

Texte intégral.

• 1 M. Blanchet , “Industrie 4.0, Nouvelle donne industrielle, nouveau modèle économique”, Outre-Terre , (...)
• 2 “L’activité économique par laquelle une personne propose ou assure à distance et par voie électron (...)

1 At the dawn of Industry 4.0, combining automation, the Internet of Things and artificial intelligence 1 , the French legislator, with European impulse, had already taken a stand in favour of the digital economy, defining electronic commerce as “the economic activity by which a person offers or ensures the supply of goods or services at a distance and by electronic means” 2 . However, if the notion of electronic commerce and distance contracts is now integrated into the French and European legislative framework – but also into usage – a new disruptive innovation is turning contractual practices upside down.

• 3 The most widely used being Ethereum, but the protocols are executable on any blockchain. See in th (...)

4 J.-Ch. Roda , Smart contacts, dumb contracts?, Dalloz IP/IT, 2018, p. 397.

• 5 Axa’s Website, « Axa se lance sur la Blockchain avec fizzy », [online] available on : https://www. (...)

2 The smart contract is in fact an execution of a computer code backed by a blockchain 3 . Known for years in the United States but not yet widely used in France 4 , it is not a simple dematerialised contract, but an automatically executed contract subject to algorithmic conditions. This means that the smart contract is a protocol for executing pre-established computer orders, and only when the predefined conditions are met will it proceed automatically and instantaneously to execution. This can be illustrated in the context of insurance contracts: Axa, an insurance company, for example, has recently introduced the Fizzy service 5 , a so-called “parametric” insurance. These insurance contracts can be taken out when purchasing a train or plane ticket, and if a delay occurs it is automatically noted by the smart contract and reimbursement is automatically carried out. In this way, smart contracts save time, but also ensure reliability, since the conditions of execution are guaranteed by the blockchain.

6 Ch. Caron , L’IA, ou le retour de HAL, Communication, Commerce Électronique , 2018, n° 5, p. 1.

• 7 B. Barraud , Le droit en datas : comment l’intelligence artificielle redessine le monde juridique, (...)

8 E. Charniak , Introduction au Deep Learning , Dunod, 2021.

• 9 Y. Le Cun , Quand la machine apprend, La révolution des neurones artificiels et de l’apprentissage (...)

10 E. Charniak , ibid.

• 11 CNIL, CNIL’s website, Intelligence artificielle, l’avis de la CNIL et ses homologues sur le futur (...)

12 CNIL, ibid.

13 CNIL, ibid.

3 In addition to smart contracts, the issue of artificial intelligence is also at the heart of the lawyer’s concerns 6 . Although both technologies are integral parts of the digital revolution that is creating Industry 4.0, they must be distinguished. Artificial intelligence is a field of technology, digital or otherwise, which aims to imitate human intelligence to assist, support or even replace humans in their tasks 7 . There are several techniques for this purpose, the best known – because it is the most powerful and recent 8 – being deep learning 9 . The latter is itself a form of machine learning 10 , which is only one of several artificial intelligence techniques. However, in view of the technical complexity and legal insufficiency, the European Union defines and classifies, in its draft regulation on artificial intelligence, several levels of risk, not in relation to the technique, but according to the risk that weighs on the users of artificial intelligence 11 . Particular attention is also paid to the personal data 12 that feeds artificial intelligence through Big Data. It is therefore possible to see that the European approach seems to be built around risk, and seeks above all to set limits, to create prohibitions; but it is also aware of the inevitable innovation and its benefits 13 .

• 14 See : B. Nicolleti , The Future of FinTech, Integrating Finance and Technology in Financial Service (...)
• 15 See : M. Corrales , M. Fenwick , H. Haapio , Legal Tech, Smart Contracts and Blockchain , Springer, 20 (...)
• 16 A. Charpentier , Big Data, GAFA et assurance, Annales des Mines – Réalités industrielles , 2020, n°  (...)
• 17 See : C. Grenier , H. Hudebine , B. Pauget , Innovation en santé : un renouvellement conceptuel et mé (...)
• 18 Ch. Larroumet , S. Bros , Les obligations, Le contrat , in : Traité de Droit civil , (dir.) Ch.  Larrou (...)
• 19 B. De Bertier-Lestrade , La bonne foi dans la réforme française des contrats, in : Le contrat dans (...)
• 20 B. Barraud , Le droit en datas : comment l’intelligence artificielle redessine le monde juridique, (...)

4 Moreover, these two technologies, smart contracts on blockchain and artificial intelligence, also have multiple applications. We can mention, among others, FinTech in the financial sector 14 , LegalTech in the legal sector 15 , InsurTech in the insurance sector 16 , and MedTech in the medical sector 17 . Although different names are used in practice, these technologies are essentially the same use: an application of digital innovations to sectors already regulated by positive law. In the context of property exchanges between people, the contract is the legal tool of reference 18 . However, if positive contract law is built around the trust of co-contractors in the rule of law – which guarantees the respect of wills and good faith 19 – the use of new digital tools transfers this trust to digital technology 20 . Whereas the law was the limit framing wills, technology becomes the guarantor of the exact will, of a precise and immutable, even mathematical will, unless the law frames the technology.

• 21 Ordonnance n° 2016-131 du 10 février 2016 portant réforme du droit des contrats, du régime général (...)
• 22 S. Lequette , ibid.  ; V. Valais, La réforme du code civil : quels enjeux pour nos contrats ?, Dallo (...)

5 Indeed, the will is necessarily a human characteristic that is different from necessity or obligation. The will is imbued with freedom, but in the context of a contractual execution by computer, it is sealed, engraved in the blockchain, and deprived of freedom, the keystone of the legal regime of contracts. Therefore, it is useful to study the impact of these technologies on contract law. The reform carried out by the ordinance of 10 February 2016 21 was intended to modernise French contract law, which had remained unchanged since the Napoleonic Code. Although it brilliantly enshrined developments in the field over two centuries 22 , it is regrettable that there was no legislative innovation anticipating digital developments, even though they were already flourishing when it was adopted. Moreover, the legislative rules governing digital technologies are also being developed under the impetus of the European legislator, particularly since the adoption of the General Data Protection Regulation (GDPR). However, unlike the European Union, which adopts a cautious approach, it must be demonstrated that digital tools can be favourable to the formation and proper performance of a contract.

6 Therefore, it is essential to show first that positive contract law, although deliberately broad and flexible, cannot correspond to the reality of augmented, intelligent, or artificial intelligence-assisted contracts (I). Consequently, it appears that a framework for these practices, although not without a certain amount of freedom, would make it possible to promote their development with a view to benefiting the company using them (II).

I. The limits of positive law

• 23 A.-C. Mansion , L’émergence des smart contracts : une future révolution juridique, in: L’impact des (...)
• 24 N. Szabo , Formalizing and Securing Relationships on Public Networks, First Monday , 1 er  sept. 1999; (...)

7 The smart contract is not a contract in the sense of French positive law 23 ; however, it is used for the same purpose. The aim of classic digital contractualisation or the smart contract is to create and regulate obligations between parties 24 . Now, it is possible to try to find a correspondence between the technique of smart contracts and artificial intelligence and contract law, but if it is possible to detect a will in the sense of the general regime during the formation of the smart contract (A), its technical execution is remote from legal considerations (B).

A. The formation of the smart contract

25 Article 1113 al. 1, French Civil Code.

26 J. Ghestin , La notion de contrat, Recueil Dalloz , 1990, chron. 147.

27 Ch. Larroumet , S. Bros , op. cit. , §87.

28 Article 1101, French Civil Code.

• 29 A.-C. Mansion , op. cit. : C.  Zolynski , Fintech – Blockchain et smart contracts : premiers regards (...)

8 “The contract is formed by the meeting of an offer and an acceptance by which the parties manifest their will to commit themselves” 25 , so in terms of positive law, it is possible to qualify the contract as soon as there is a manifestation of the will to contract. The will was set up by the classical theory as the basis of the contract 26 , and if modern theories have been able to question its fundamental place, it appears that ‘there is no contract without the will to contract and that, consequently, it is difficult to deny that the contract is based on an act of the will’ 27 . The definition adopted by French positive law enshrines the will as a central notion 28 . However, it is possible to ask at what point the smart contract is really formed, or even if it is really a contract in the sense of positive law 29 . Indeed, several stages are necessary for its creation.

• 30 C. Barreau , La régulation des smart contracts et les smart contracts des régulateurs, Annales des (...)
• 31 Which one could be retained for an AFNOR standard: CE, 6 e  Ch., 28 juill. 2017, M.   A.   c. Ministère (...)

32 A.-C. Mansion , op. cit.

9 First, the computer code that automatically executes the contract must be written by human intervention 30 . This step may in fact be like the negotiation of a classical contract, only the language differs. However, since it is a computer executable program, it is also possible to question whether copyright can be granted. Article L. 112-2, 13° of the Intellectual Property Code expressly mentions software as being protectable by copyright, and moreover the fact that this software constitutes a legal act does not seem to oppose the qualification of a work 31 , provided that it is original. This qualification would then place the designer of the execution program in a position of strength. However, the manifestation of the parties’ will is expressed when both parties accept the conditions of execution of the smart contract previously established or negotiated, even if it was created by one of the parties or a third party. In this sense, the parties must expressly accept the conditions of performance of the smart contract 32 . It appears then that in a legal sense the contract is formed at the moment when the parties have expressed their consent to the terms of the contract, and this provision seems to be appropriate for both classical and smart contracts.

• 33 M. Danis , Ch. Bouffier , Th. Feigean, L’intelligence artificielle appliquée au secteur de la financ (...)
• 34 Although artificial intelligence is present in the design of smart contracts, it has not, to date, (...)

10 However, it is also appropriate to question the relevance of positive law on the formation of the contract when it is assisted by artificial intelligence. Indeed, the focus is no longer really on the formation of the contract but more on the negotiation 33 . When an artificial intelligence is involved in the making of the contract, it is indeed possible to validly question the principle of freedom of negotiation set out in Article 1112 paragraph 1 of the Civil Code. Assuming that an artificial intelligence could draft – or program – smart contracts 34 , the negotiations essential to the contract would be deprived of part of their substance. However, the actions of artificial intelligence save time; it is possible to imagine clauses automatically drafted according to the contractual habits of the parties or personalised performance conditions with a high probability of agreeing with the co-contractors. In this sense, it is possible to say that contractual freedom lies in the choice of whether or not to accept the conditions proposed by artificial intelligence. It appears then that positive contract law can be adapted to the integration of new digital technologies in the formation and negotiation of contracts. However, it is necessary to determine whether it can be adapted to the execution of the contract.

B. The execution of the smart contract

• 35 H. Claret , Interprétation des contrats d’assurance et droit de la consommation, Recueil Dalloz , 20 (...)

36 Article 1188, French Civil Code . ; Article 12, al. 3, French Civil Procedure Code .

11 The smart contract is the computer code that automatically executes the wishes of the parties. A first point of contention can be raised: language. Indeed, while legal language allows a certain margin of interpretation 35 , computer language, which is mathematical, is marked by formality. Therefore, a first argument in favour of smart contracts emerges: mathematical language succeeds where legal language fails in terms of precision. Smart contracts thus offer no room for subjective interpretation. But this mathematical precision is not necessarily an advantage; legal language allows the parties, or if necessary the judge 36 , to reinterpret the contract according to the case in question in order to reflect their actual wishes. The use of smart contracts makes it possible to express the will of the parties at the time of the conclusion of the contract in an inescapable way, but it does not make it possible to follow its evolution. This makes it difficult to renegotiate the terms, and the performance of the contract remains subject to predefined mathematical rules: this can be seen as an impediment to the fundamental principle of contractual freedom.

• 37 See on the example of the execution of a shareholders’ agreement in case of a sale of shares: J.-C (...)

38 See : D. Houtcieff , Droit des contrats , 6 e  éd, Bruylant, 2021, spéc. §132-10.

• 39 M. Mekki , Blockchain , l’exemple des smart contracts , Entre innovation et précaution, mai 2018 [onl (...)
• 40 L. Aynès , La confiance en droit privé des contrats – Rapport de synthèse, in : La confiance en dro (...)

12 However, the smart contract is not a tool without advantages; on the contrary, there are quite a few. The execution of the contract is automatic 37 , reliable 38 and instantaneous. The idea of an automatically executed contract is not new, it existed before the creation of the blockchain, but the advent of smart contracts allows, above all, the consolidation of a bond of trust between two or more individuals, which was previously non-existent 39 . However, if the securing of contractual exchanges by the blockchain opens up new economic possibilities, it calls into question the legal normative framework. Indeed, trust is a central notion in contract law 40 . It is therefore legitimate to ask whether it is possible to qualify smart contracts as contracts when the trust is artificial and based on technology and not on the co-contractor or the law.

41 M. Danis et al ., op. cit.

13 Moreover, the relevance of positive law is even more questionable when artificial intelligence is involved in the actual performance of the contract. Indeed, if one accepts the existence of a smart contract that is not simply automated but assisted by artificial intelligence in its performance 41 , it is possible to imagine a contract whose content is the collection of data so that they can be processed. An execution error attributable to the algorithm would lead to a fault in the execution of the contract. To guarantee good faith performance by the co-contracting parties, it is essential to provide for contractual adaptations, particularly regarding liability in the event of an error attributable to the artificial intelligence, but also for ethical provisions concerning its use. While positive law provides that contracts must be executed in good faith, it is possible to deny the existence of such a contract within an algorithm, unless the legal standard can provide for certain precautions to guarantee good faith.

II. The search for a forward-looking law

14 Smart contracts or contracts assisted by artificial intelligence are business tools that escape the legal regulation of contract law. Therefore, it seems fundamental to seek a forward-looking law that allows for regulation favourable to the development of new digital contractual tools (A). However, this regulation must be marked by a certain liberty to allow the development of innovation (B).

A. The need for regulation

• 42 E. Orenna , préf. in : G. Babinet , Big data. Penser l’homme et le monde autrement , Le passeur, 2015 (...)

43 A.-C. Mansion , op. cit.  ; N.  Szabo , op. cit.

• 44 C. Boismain , Quelques réflexions sur les contrats intelligents (smart contracts), LPA , 2021, n° 15 (...)

45 C. Boismain , op. cit.  ; A.-C.  Mansion , op. cit. 

46 This is the principle, but there are exceptions: Article 1103, French Civil Code.

47 H. Claret , op. cit.  ; M.  Lamoureux , op. cit.

15 First, the smart contract should be understood as an augmented and non-automated contract: in this sense, it is essential to argue in favour of a specific regulation 42 . Indeed, the smart contract is a digital solution that offers more possibilities than the classic contract or the electronic contract. It allows solutions to be set up according to external events and in an automated manner. It is in fact a real living contract that offers several possibilities and that can be adapted according to external hazards, as long as the computer code that executes it so provides 43 . However, as it has no legal recognition, it is inappropriate to speak of contracts in the legal sense, as the formation of the smart contract differs from the formation of the classic or electronic contract 44 ; there is also no provision for any form of legal withdrawal – unless it is provided for in the computer code of the smart contract itself 45 – as the commitment of the parties is definitively recorded in the blockchain. In this sense, it is desirable that the legal standard should regulate the formation of the smart contract, in order to avoid defects in consent, but also in its execution. Indeed, the execution may be subject to hazards external to the contract: although certified by the blockchain, the realisation of these hazards will automatically and definitively execute the contract, without taking into consideration events of force majeure. In short, the smart contract offers a certain security by making it impossible to breach the contract 46 , but at the detriment of the flexibility that is allowed by the legal standards governing traditional or electronic contracts. This argument can be put forward in particular with regard to consumer law 47 , which aims to protect consumers from abuse by professionals. Without a legal framework, the smart contract cannot be a technological advance in that it would jeopardise the protections granted to certain parties in the conclusion of contracts.

• 48 H. Jacquemin et J.-B. Hubin , Aspects contractuels et de responsabilité civile en matière d’intelli (...)
• 49 P. De Filippi , S. Hassan , Blockchain technology as a Regulatory Technology: From Code is Law to La (...)
• 50 I. A. Omar et al., Ensuring protocol compliance and data transparency in clinical trials using Blo (...)
• 51 CNIL, Intelligence artificielle, l’avis de la CNIL et ses homologues sur le futur règlement europé (...)

16 The use of artificial intelligence in contractual relations is not new either. In the case of an online purchase, the contract is concluded in an automated manner without the need for any intervention on the company’s side. Similarly, intelligent stock management software can place an order automatically without a physical person having to give the order 48 . The use of artificial intelligence in contractual techniques can to some extent be considered as progress, but it increases the possibility of defects in consent and makes their proof more difficult 49 . Indeed, it is difficult to prove an error or fraud when it has been committed by an artificial intelligence that acts automatically, or even autonomously. This raises questions of liability and the burden of proof, which complicate contract law. The solution, however, lies in a clear legislative framework providing for rules of compliance by design 50 . It is in this sense that the future European regulation on artificial intelligence 51 can be seen as a considerable step forward.

• 52 N. Martial-Braz , Le droit des contrats à l’épreuve des géants d’Internet, in : L’effectivité du dr (...)

17 The question of smart contract law also raises compliance prerogatives. There are no rules in French positive law imposing the regulation of artificial intelligence or the use of smart contracts, and even if there were, the law is only territorial in scope and may seem powerless in the face of the internationality of digital technology 52 . The regulation of digital tools cannot therefore come exclusively from the national legislator, but must come from a supranational entity, such as the European Union, which will be able to impose compliance. In this sense, it is essential that companies be able to apply compliance measures to a regulation that protects the parties to the contract and guarantees a certain legal certainty. Nevertheless, these measures should not be too restrictive in order to avoid inhibiting innovation.

B. The need for liberty

53 P. De Fillipi , S. Hassan , op. cit.

• 54 Loi n° 2004-575 du 21 juin 2004, pour la confiance dans l’économie numérique, Article 14, JORF , 22 (...)

18 A solution that would allow some regulation of the new contractual tools without inhibiting their development would be to ensure that the provisions protecting the parties are included in smart contracts or contracts developed by artificial intelligence in an automatic manner 53 . This provision would not be contrary to the principle of contractual freedom, but it would allow public policy provisions to be inserted into smart contracts or computer programs governing the formation and performance of the contract. To do this, it is necessary to enshrine a law of augmented contracts, or contracts assisted by digital technologies. It is not a question of creating a sui generis right but, as in electronic contracts 54 , only of saying that the public policy provisions of the Civil Code relating to contract law must be able to be applied to augmented or smart contracts.

• 55 B. Deffains , Le monde du droit face à la transformation numérique, Pouvoirs , 2019, n° 170, p. 43-5 (...)

56 B. Deffains , ibid.

19 However, this solution poses technical difficulties, since it amounts to regulating by law a contractual technique which is based on digital technology. In this sense, it is important that the computer developer and the engineer be associated with the lawyer in the creation process. This implies the advent of new, hitherto unknown professions 55 such as legal engineer or legal developer. The advent of these professions also marks an evolution in the role of the lawyer within the company. Indeed, the lawyer will no longer only have to be able to master the law but also the technology that will implement it in the future. The advent of artificial intelligence and digital techniques, such as the blockchain, show a clear evolution in the purpose of law 56 .

• 57 See : E. Chenut , D. Quercioli , Le numérique au service de l’humain, ou comment promouvoir un usage (...)

20 In short, the solution aimed at regulating the technique that executes the law will only really be put in place when the developers and users of these techniques have a dual competence. However, in order to give impetus to these innovations, which have multiple advantages, it is essential that a normative framework pushes the lawyer to take a closer interest in them 57 . Legal developments can be devoted to each legal aspect that has its legal equivalent, but the digital world no longer spares any sector; it is a virtual mirror of real society, or even the reality of tomorrow. In this sense, it is essential to promote digital rights and to encourage their development, which will be useful to companies, by means of a law of innovation.

1 M. Blanchet , “Industrie 4.0, Nouvelle donne industrielle, nouveau modèle économique”, Outre-Terre , 2016, n° 46, p. 62-85, §1.

2 “L’activité économique par laquelle une personne propose ou assure à distance et par voie électronique la fourniture de biens ou de services”: Loi n° 2004-575 du 21 juin 2004, pour la confiance dans l’économie numérique, Article 14, JORF , 22 nd  June 2004, n° 0143 text 2.

3 The most widely used being Ethereum, but the protocols are executable on any blockchain. See in this sense : S. Tikhomirov et al., SmartCheck: static analysis of ethereum smart contracts, in: Proceedings of the 1st International Workshop on Emerging trends in Software Engineering for Blockchain - WETSEB’18, Association for Computing Machinery , 2018, pp. 9-16; on the notion of blockchain: Th. Douville, Blockchain et protection des données à caractère personnel, AJ Contrat , 2019, p. 316; J.  Gossa , Les blockchains et smart contracts pour les juristes, Dalloz IP/IT , 2018, p. 393; M. M ekki , Les mystères de la blockchain, Dalloz IP/IT , 2017, p. 2160.

5 Axa’s Website, « Axa se lance sur la Blockchain avec fizzy », [online] available on : https://www.axa.com/fr/magazine/axa-se-lance-sur-la-blockchain-avec-fizzy (accessed on 1st September 2021).

7 B. Barraud , Le droit en datas : comment l’intelligence artificielle redessine le monde juridique, Revue Lamy Droit de l’Immatériel , 2019, n° 164.

9 Y. Le Cun , Quand la machine apprend, La révolution des neurones artificiels et de l’apprentissage profond , Éd. Odile Jacob, 2019, spéc. p. 225 et s.

11 CNIL, CNIL’s website, Intelligence artificielle, l’avis de la CNIL et ses homologues sur le futur règlement européen, 8 juill. 2021, [online] available on: https://www.cnil.fr/fr/intelligence-artificielle-lavis-de-la-cnil-et-de-ses-homologues-sur-le-futur-reglement-europeen accessed on 1 st  Sept. 2021).

14 See : B. Nicolleti , The Future of FinTech, Integrating Finance and Technology in Financial Services , Springer, 2017.

15 See : M. Corrales , M. Fenwick , H. Haapio , Legal Tech, Smart Contracts and Blockchain , Springer, 2019.

16 A. Charpentier , Big Data, GAFA et assurance, Annales des Mines – Réalités industrielles , 2020, n° 1, p. 53-57.

17 See : C. Grenier , H. Hudebine , B. Pauget , Innovation en santé : un renouvellement conceptuel et méthodologique pour transformer durablement le champ de la santé, Innovations , 2021, n° 65, p. 5-19.

18 Ch. Larroumet , S. Bros , Les obligations, Le contrat , in : Traité de Droit civil , (dir.) Ch.  Larroumet , tome 3, Economica, 9 th  ed., 2018, §3 et s.

19 B. De Bertier-Lestrade , La bonne foi dans la réforme française des contrats, in : Le contrat dans tous ses états, (dir.) C. Le Gallou et A.  Marmisse-D’abadie   D’arrast , PUT1, Actes de colloque de l’IFR, n° 41, p. 141-160.

20 B. Barraud , Le droit en datas : comment l’intelligence artificielle redessine le monde juridique, Revue Lamy Droit de l’Immatériel , 2019, n° 164.

21 Ordonnance n° 2016-131 du 10 février 2016 portant réforme du droit des contrats, du régime général et de la preuve des obligations, JORF , n° 0035, du 11 th  Feb. 2016, text n° 26 ; S.  Lequette , La réforme du droit commun des contrats et contrats d’intérêt commun, Recueil Dalloz , 2016, p. 1148.

22 S. Lequette , ibid.  ; V. Valais, La réforme du code civil : quels enjeux pour nos contrats ?, Dalloz IP/IT , 2016, p. 229.

23 A.-C. Mansion , L’émergence des smart contracts : une future révolution juridique, in: L’impact des nouvelles technologies sur le droit et ses acteurs, Colloque ADDCDA, Annales de l’Université Toulouse 1 Capitole , tome LIX, 2019, I, p. 382 ; Y.  Cohen-Hadria , Blockchain  : révolution ou évolution ?, Dalloz IP/IT , 2016, p. 537.

24 N. Szabo , Formalizing and Securing Relationships on Public Networks, First Monday , 1 er  sept. 1999; F.  Terré , Ph.  Simler , Y.  Lequette , F.  Chénedé , Droit civil, Les obligations , Dalloz, 12 e  éd., 2019, §17.

29 A.-C. Mansion , op. cit. : C.  Zolynski , Fintech – Blockchain et smart contracts : premiers regards sur une technologie disruptive, Revue de Droit bancaire et financier , 2017, n° 8.

30 C. Barreau , La régulation des smart contracts et les smart contracts des régulateurs, Annales des Mines – Réalités industrielles , 2017, n° 3, p. 74-76.

31 Which one could be retained for an AFNOR standard: CE, 6 e  Ch., 28 juill. 2017, M.   A.   c. Ministère de l’Environnement  : Communication, Commerce Électronique , 2017, com. 78, note Caron ; Propriétés Intellectuelles 2018, n° 66, p. 59 obs. Bruguière ; but also for a legal treaty, Cass. 1 ère  Civ., 9 janv. 1996 : Revue Internationale du Droit d’Auteur, 1996, n° 169, 331, obs. Kéréver.

33 M. Danis , Ch. Bouffier , Th. Feigean, L’intelligence artificielle appliquée au secteur de la finance : enjeux contractuels et responsabilités, Annales des Mines – Réalités industrielles , 2019, 1, p. 65-68 ; Ph.  Mathieu , M.-H.  Verrons , GeNCA : Un modèle général de négociation des contrats, Revue des Sciences et Technologies de l’Information – série Revue d’Intelligence Artificielle , 2005, vol. 19, n° 6, p. 837-884.

34 Although artificial intelligence is present in the design of smart contracts, it has not, to date, been a content creator: C. Barreau, op.   cit .

35 H. Claret , Interprétation des contrats d’assurance et droit de la consommation, Recueil Dalloz , 2003, p. 2600 ; M. Lamoureux , L’interprétation des contrats de consommation, Recueil Dalloz , 2006, p. 2848.

37 See on the example of the execution of a shareholders’ agreement in case of a sale of shares: J.-Ch. Roda , Smart contacts, dumb contracts?, Dalloz IP/IT , 2018, p. 397.

39 M. Mekki , Blockchain , l’exemple des smart contracts , Entre innovation et précaution, mai 2018 [online] available on: https://lesconferences.openum.ca/files/sites/97/2018/05/Smart-contracts.pdf (accessed on 1 st  Sept. 2021).

40 L. Aynès , La confiance en droit privé des contrats – Rapport de synthèse, in : La confiance en droit privé des contrats , (dir.) V.-L.  Bénabou et M.  Chagny , 2008, Dalloz, p. 151 ; V.  Edel , La confiance en droit des contrats , (dir.) R.  Cabrillac , thèse dactyl. Montpellier I, 2006.

42 E. Orenna , préf. in : G. Babinet , Big data. Penser l’homme et le monde autrement , Le passeur, 2015, p. 6.

44 C. Boismain , Quelques réflexions sur les contrats intelligents (smart contracts), LPA , 2021, n° 158, p. 6.

48 H. Jacquemin et J.-B. Hubin , Aspects contractuels et de responsabilité civile en matière d’intelligence artificielle, in : L’intelligence artificielle et le droit , (dir.) H.  Jacquemin et A.  De Streel , Larcier, 2017, p. 104, §30.

49 P. De Filippi , S. Hassan , Blockchain technology as a Regulatory Technology: From Code is Law to Law is Code, First Monday , 8 janv. 2018.

50 I. A. Omar et al., Ensuring protocol compliance and data transparency in clinical trials using Blockchain smart contracts, BMC Medical Research Technology , 2020, n° 20, 224.

51 CNIL, Intelligence artificielle, l’avis de la CNIL et ses homologues sur le futur règlement européen, op. cit.

52 N. Martial-Braz , Le droit des contrats à l’épreuve des géants d’Internet, in : L’effectivité du droit face à la puissance des géants de l’Internet , vol. 1, (dir.) M.  Behar-Touchais , IRJS éditions, tome 63, p. 61.

54 Loi n° 2004-575 du 21 juin 2004, pour la confiance dans l’économie numérique, Article 14, JORF , 22 juin 2004, n° 0143 texte 2.

55 B. Deffains , Le monde du droit face à la transformation numérique, Pouvoirs , 2019, n° 170, p. 43-58.

57 See : E. Chenut , D. Quercioli , Le numérique au service de l’humain, ou comment promouvoir un usage facteur d’émancipation individuelle collective, in : Numérique, action publique et démocratie , (dir.) Ph.  Bance , J.  Fournier , PURH, 2021, p. 237 et s.

Doctorant, Université Toulouse 1 Capitole, CDA

Le texte et les autres éléments (illustrations, fichiers annexes importés) sont sous Licence OpenEdition Books , sauf mention contraire.

Vérifiez si votre institution a déjà acquis ce livre : authentifiez-vous à OpenEdition Freemium for Books. Vous pouvez suggérer à votre bibliothèque/établissement d’acquérir un ou plusieurs livres publié(s) sur OpenEdition Books. N'hésitez pas à lui indiquer nos coordonnées : OpenEdition - Service Freemium [email protected] 22 rue John Maynard Keynes Bat. C - 13013 Marseille France Vous pouvez également nous indiquer à l'aide du formulaire suivant les coordonnées de votre institution ou de votre bibliothèque afin que nous les contactions pour leur suggérer l’achat de ce livre.

Merci, nous transmettrons rapidement votre demande à votre bibliothèque.

Volume papier

Référence électronique du chapitre, référence électronique du livre, collez le code html suivant pour intégrer ce livre sur votre site..

OpenEdition est un portail de ressources électroniques en sciences humaines et sociales.

• OpenEdition Journals
• OpenEdition Books
• OpenEdition Freemium
• Mentions légales
• Politique de confidentialité
• Gestion des cookies
• Signaler un problème

Vous allez être redirigé vers OpenEdition Search

Artificial Intelligence and Law

• Cyber Law Subject-wise Law Notes
• July 6, 2021

‘It is unworthy of excellent men to lose hours like slaves in the labor of calculation which could safely be relegated to anyone else if machines were used.’

– Gottfried Wilhelm   Leibniz

Law touches every topic of the world. Today, the legal services market is one of the largest globally but extremely under-digitized. The field is based on traditional concepts, thus, slow to adapt new technologies and methods.

In this essay, the prime focus is on the expected development in practice of law through artificial intelligence.

In nearly all of the social sciences, law may come closest to a system of formal logic. We focus on rulings which set forward the principles deriving from precedents. These precedents are, then, applied on different facts at hand to reach a conclusion accordingly. This logic-oriented methodology is exactly the type of activity to which machine intelligence can fruitfully be applied. Till now, MS Word and mails were the most commonly used tools in legal work while there seems to be a whole opportunity to make this process completely automatic.

What is AI?

Artificial Intelligence [AI] is a computer system able to perform tasks which normally require human intelligence. These systems are mainly powered by machine learning, some by deep learning and rules. This comes with learning which involves garnering the rules and information for using the data. It has become very popular and necessary due to data-based service industries like telecommunication, insurance, banking, etc. including law.

Today’s AI systems are not intelligent thinking machines in any meaningful sense though they produce useful and intelligent results without using intelligence. These systems do this by detecting patterns in different data and using knowledge, rules, and information specifically encoded in forms able to be processed by computers. Through these computational methods, AI systems produce commendable results on complex tasks that, if done by humans, require cognition. However, the systems do it with the help of computational methods which are not at all in accordance with the human mind.

AI applications today are of two forms i.e. “narrow AI” or artificial specialized intelligence (ASI) and “general AI” or artificial general intelligence (AGI).

The “narrow AI” aims to solve specific problems or take actions within a limited set of parameters such as communicating with a device to book film tickets or pay a bill or listen to GPS directions, Using Google translate service. This type of AI appears to be smart but it lacks functions other than what is defined. To simplify, it has zero self-awareness. Artificial Intelligence can facilitate a machine for copying the cognitive functions which are naturally ingrained in human beings to associate with other beings such as problem-solving or grasping new data. 

The latest AI software is not too flexible to adapt in terms of switching unrelated activities, i.e. from one to another, unrelated activities. It would be a fault on our part to assume that just because AI knocks the human thought process in some tough game, it will necessarily lead to the automation of other difficult tasks like creative legal argumentation or problem solving. Simply put, contemporary AI tends to work best for activities where there are underlying patterns, rules, definitive right answers, and semi-formal or formal structures that make up the process and it works poorly in conceptual, abstract, value-laden, open-ended, policy or judgment-oriented activities. 

Why Linked to Law?

In recent times, AI has changed the working methods of multiple industries by being installed at a highly effective scale. The Indian legal sector has seen very little of this innovation as the legal practitioners still rely on the methods designed ages ago. The system is vast and constantly changing and with the use of Artificial Intelligence, lawyers can get unparalleled insight into the legal domain within seconds.

Currently we get all of our legal research done by a significant number of man-hours which eventually reduces the profit-making ability of a firm. However, with AI, the entire legal fraternity can balance the expenditure required for the research work by using a uniform method for the same.

In 2019, it was announced that students at University College London and the University of Sheffield had successfully developed artificial intelligence software that can predict the outcome of human rights cases by analysing previous court judgments. The story sounded like science fiction but the AI software showed an astonishing 79% accuracy. The role of technology in law is influenced by how a client wants the problem to be solved as he needs the legal advice of his lawyer. The practitioners of law perform various legal tasks, which involves client counseling, drafting contracts, petitions and other documents, practicing in court, and many more. AI is not magic, rather it produces results by looking out for different patterns or rules (heuristic proxies) to make decisions on its own. Neither is it good with understanding meanings or handling totally undeveloped or open-ended tasks. AI’s successful tasks involve highly structured data with some clear   answers and patterns to distinguish the suitable solution. Knowing AI’s flaws and pros help us to understand the impact it is going to have on different fields of law. 

There are certain startups initiated by human minds as they tried to move it beyond just MS Word and mails.  For instance, these programs use Q/A method to compile all the data given by the client to form a perfect contract according to their needs. 

How Does It Help? 

AI is said to have a great scope for Indian Legal fraternity as the amalgamation of two will lead to tremendous development of both in the near future. Currently, these fields are proved to be useful in law:

1.     Contract reviewing   – Lawyers manually review, edit and exchange red-lined documents and the process is lengthy, delays deals and impedes companies’ business objectives. Mistakes due to human error are common as the contracts can be thousands of pages long. To do all these tedious works with due diligence, AI is proving to be helpful and time effective.

2.   Legal Analytics  – AI helps digging into numerous past case laws, judgments, precedents, rules, obiter dictums, etc. to analyse and use the same in present cases.

3.  Outcome Prediction  – It also predicts the probable outcome of different pending cases based on previous judgments,facts and/or precedents. As these predictions become more accurate, they will have a major impact on the practice of law.

4.   Legal Research  – The machine intelligence is making inroads in legal research effectively. Legal research was historically a manual process, but with the advent of software and personal computing, it has gone digital. Lawyers usually conduct research using websites like LexisNexis, SCC and Manupatra.

5.   Automation of Documentation  – By just submitting the required documents which you wish to incorporate in your legal document get your documents ready within minutes.

6.   Intellectual Property  – People can search for the uniqueness of their logos, trademarks, patents, copyrights, etc as well as check the registration status without stepping out of their homes. 

Replacement for Jobs Or A Support For Better Performance?

There is a burning question among the lawyers i.e. whether the introduction of Artificial Intelligence in the law and legal sector would replace the lawyers and analysts OR the AI-based solutions would lead to increased productivity for them? 

As already discussed above, the legal field has been introduced to various developments by AI leading to increased efficiency of legal practitioners by giving astonishing results in issues like trademark searches, contract analysis, etc. mainly focusing on legal research. When we see it, these remarkable technological developments do not appear to leave very much room for human lawyers to get involved. However, it is also being used to make lawyers’ lives easier and more flexible.

Technology gives legal professionals the opportunity to work from home. Some websites connect some experienced freelance lawyers bidding for a range of legal work. All these advances are not only allowing lawyers within law firms to work more efficiently, but also to change the ways lawyers interact with their clientele. This, therefore, helps them to move forward from what was a traditional method to a modern one for their firms.

Furthermore, these technological changes are very dubious to make these actual and tangible lawyers unessential any sooner.  While AI has been proven a great resource to us, there is a limit to what formulaic and logical programmes it can achieve.

To quote Leibniz, one of the grandfathers of AI and a lawyer: ‘It is unworthy of excellent men to lose hours like slaves in the labor of calculation which could safely be relegated to anyone else if machines were used.’

In 1673, he presented the machine for four arithmetic operations in the UK. Leibniz says ‘The only way to correct our reasoning is to make them as tangible as the mathematicians’ so that we can find our error at a glance, and when there are disagreements between people, let’s calculate and see who is right!’

The legal profession basically works upon cognitive skills like analysing, decision making, and representation. These skills cannot be taken over by AI, hence, can never be automated. Moreover, clients value highly the social and public skills of their lawyers. Employers will indeed expect legal candidates to be well-experienced in modern technology but with the thought of proficient manual labour as well. A candidate will still be asked to interact with their patrons by using their experience in public dealing and knowledge of law for finding solutions to all legal problems.  

Conclusion 

It is likely that technological innovation will continue to reflect clients’ needs. AI is certainly set to change the way lawyers work and will hopefully make day-to-day tasks like drafting and legal research far more efficient and cost-effective. However, clients will ultimately still want an experienced legal adviser in their corner who can use new technology to their advantage while maintaining a unique inter-personal relationship. Concluding the same, I would like to focus on the points AI will prove itself helpful in:

• The way clients are serviced and treated in the future. Firms would approach with ideas authentic and never known before 
• Shifting the focus of Firms will shift from revenue to higher profits.
• Making Technology the new trend i.e. It is a good time for start-ups to emerge.

Author- Bhoomika Sharma (Amity Law School, Delhi)

You might like

Subsidiary Rules of Interpretation

Closure in Labour Law

Parliamentary Sovereignty in India

Leave a reply cancel reply.

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

Name  *

Email  *

Add Comment  *

This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.

Post Comment

Artificial intelligence as law

Presidential address to the seventeenth international conference on artificial intelligence and law

• Review Article
• Open access
• Published: 14 May 2020
• volume  28 ,  pages 181–206 ( 2020 )

You have full access to this open access article

• Bart Verheij 1  

14k Accesses

23 Citations

13 Altmetric

Explore all metrics

Cite this article

Information technology is so ubiquitous and AI’s progress so inspiring that also legal professionals experience its benefits and have high expectations. At the same time, the powers of AI have been rising so strongly that it is no longer obvious that AI applications (whether in the law or elsewhere) help promoting a good society; in fact they are sometimes harmful. Hence many argue that safeguards are needed for AI to be trustworthy, social, responsible, humane, ethical. In short: AI should be good for us. But how to establish proper safeguards for AI? One strong answer readily available is: consider the problems and solutions studied in AI & Law. AI & Law has worked on the design of social, explainable, responsible AI aligned with human values for decades already, AI & Law addresses the hardest problems across the breadth of AI (in reasoning, knowledge, learning and language), and AI & Law inspires new solutions (argumentation, schemes and norms, rules and cases, interpretation). It is argued that the study of AI as Law supports the development of an AI that is good for us, making AI & Law more relevant than ever.

Avoid common mistakes on your manuscript.

1 Introduction

It is my pleasure to speak to you today Footnote 1 on Artificial Intelligence and Law, a topic that I have already loved for so long—and I guess many of you too—and that today is in the center of attention.

It is not a new thing that technological innovation in the law has attracted a lot of attention. For instance, think of an innovation brought to us by the French 18th century freemason Joseph-Ignace Guillotin: the guillotine. Many people gathered at the Nieuwmarkt, Amsterdam, when it was first used in the Netherlands in 1812 (Fig.  1 , left). The guillotine was thought of as a humane technology, since the machine guaranteed an instant and painless death.

Technological innovation in the law in the past (left) and in the future? (right). Left: Guillotine at the Nieuwmarkt in Amsterdam, 1812 (Rijksmuseum RP-P-OB-87.033, anonymous). Right: TV series Futurama, judge 723 ( futurama.fandom.com/wiki/Judge_723 )

And then a contemporary technological innovation that attracts a lot of attention: the self-driving car that can follow basic traffic rules by itself, so in that sense is an example of normware, an artificial system with embedded norms. In a recent news article, Footnote 2 the story is reported that a drunk driver in Meppel in my province Drenthe in the Netherlands was driving his self-driving car. Well, he was riding his car, as the police discovered that he was tailing a truck, while sleeping behind the wheel, his car in autopilot mode. His driver’s licence has been withdrawn.

And indeed technological innovation in AI is spectacular, think only of the automatically translated headline ‘Drunken Meppeler sleeps on the highway’, perhaps not perfect, but enough for understanding what is meant. Innovation in AI is going so fast that many people have become very enthusiastic about what is possible. For instance, a recent news item reports that Estonia is planning to use AI for automatic decision making in the law. Footnote 3 It brings back the old fears for robot judges (Fig.  1 , right).

Contrast here how legal data enters the legal system in France where it is since recently no longer allowed to use data to evaluate or predict the behavior of individual judges:

LOI n \(^{\mathbf{o}}\) 2019-222 du 23 mars 2019 de programmation 2018–2022 et de réforme pour la justice (1)—Article 33 Les données d’identité des magistrats et des membres du greffe ne peuvent faire l’objet d’une réutilisation ayant pour objet ou pour effet d’évaluer, d’analyser, de comparer ou de prédire leurs pratiques professionnelles réelles ou supposées. [The identity data of magistrates and members of the registry cannot be reused with the purpose or effect of evaluating, analyzing, comparing or predicting their actual or alleged professional practices.]

The fears are real, as the fake news and privacy disasters that are happening show. Even the big tech companies are considering significant changes, such as a data diet. Footnote 4 But no one knows whether that is because of a concern for the people’s privacy or out of fear for more regulation hurting their market dominance. Anyway, in China privacy is thought of very differently. Figure  2 shows an automatically identified car of which it is automatically decided that it is breaching traffic law—see the red box around it. And indeed with both a car and pedestrians on the zebra crossing something is going wrong. Just this weekend the newspaper reported about how the Chinese public thinks of their social scoring system. Footnote 5 It seems that the Chinese emphasise the advantages of the scoring system, as a tool against crimes and misbehavior.

A car breaching traffic law, automatically identified

Against this background of the benefits and risks of contemporary AI, the AI community in the Netherlands has presented a manifesto Footnote 6 emphasising what is needed: an AI that is aligned with human values and society. In Fig.  3 , key fields of research in AI are listed in rows, and in columns three key challenges are shown: first, AI should be social, and should allow for sensible interaction with humans; second, AI should be explainable, such that black box algorithms trained on data are made transparent by providing justifying explanations; and, third, AI should be responsible, in particular AI should be guided by the rules, norms, laws of society.

Artificial Intelligence Grid: foundational areas and multidisciplinary challenges (source: Dutch AI Manifesto \(^{6}\) )

Also elsewhere there is more and more awareness of the need for a good, humane AI. For instance, the CLAIRE Confederation of Laboratories for AI Research in Europe Footnote 7 uses the slogan

Excellence across all of AI For all of Europe With a Human-Centered Focus.

In other words, this emerging network advertises a strong European AI with social, explainable, responsible AI at its core.

And now a key point for today: AI & Law has been doing this all along. At least since the start of its primary institutions—the biennial conference ICAIL (started in 1987 by IAAIL), Footnote 8 the annual conference JURIX (started in 1988) Footnote 9 and the journal Artificial Intelligence & Law (in 1992)—, we have been working on good AI. In other words, AI & Law has worked on the design of socially aware, explainable, responsible AI for decades already. One can say that what is needed in AI today is to do AI as we do law.

2 Legal technology today

But before explaining how that could go let us look a bit at the current state of legal technology, for things are very different when compared to the start of the field of AI & Law.

For one thing, all branches of government now use legal technology to make information accessible for the public and to provide services as directly and easily as possible. For instance, a Dutch government website Footnote 10 provides access to laws, regulations and treaties valid in the Netherlands. The Dutch public prosecution provides an online knowledge-based system that gives access to fines and punishments in all kinds of offenses. Footnote 11 There you can for instance find out what happens when the police catch you with an amount of marihuana between 5 and 30 grams. In the Netherlands, you have to pay 75 euros, and there is a note: also the drugs will be taken away from you. Indeed in the Netherlands all branches of government have online presence, as there is a website that gives access to information about the Dutch judicial system, including access to many decisions. Footnote 12

An especially good example of successful legal technology is provided by the government’s income tax services. Footnote 13 In the Netherlands, filling out your annual tax form has become very simple. The software is good, it is easy to use, and best of all: in these days of big interconnected data much of what you need to fill in is already fillled in for you. Your salary, bank accounts, savings, mortgage interest paid, the value of your house, it is all already there when you log in. In certain cases the tool even leaves room for some mild tax evasion—or tax optimisation if you like—since by playing with some settings a married couple can make sure that one partner has to pay just below the minimal amount that will in fact be collected, which can save about 40 euros.

One might think that such legal tech systems are now normal, but that is far from true. Many countries struggle with developing proper legal tech at the government level. One issue is that the design of complex systems is notoriously hard, and this is already true without very advanced AI.

Also the Netherlands has had its striking failures. A scary example is the Dutch project to streamline the IT support of population registers. One would say a doable project, just databases with names, birth dates, marriages, addresses and the like. The project was a complete failure. Footnote 14 After burning 90 million euros, the responsible minister—by the way earlier in his career a well-recognized scientist—had to pull the plug. Today all local governments are still using their own systems.

Still legal tech is booming, and focuses on many different styles of work. The classification used by the tech index maintained by the CodeX center for legal informatics at Stanford university distinguishes nine categories (Marketplace, Document Automation, Practice Management, Legal Research, Legal Education, Online Dispute Resolution, E-Discovery, Analytics and Compliance). Footnote 15 It currently lists more than a 1000 legal tech oriented companies.

And on the internet I found a promising graph about how the market for legal technology will develop. Now it is worth already a couple of 100s of millions of dollars, but in a few years time that will have risen to 1.2 billion dollars—according to that particular prediction. I leave it to you to assess what such a prediction really means, but we can be curious and hopeful while following how the market will actually develop.

So legal tech clearly exists, in fact is widespread. But is it AI, in the sense of AI as discussed at academic conferences? Most of it not really. Most of what we see that is successful in legal tech is not really AI. But there are examples.

I don’t know about you, but I consider the tax system just discussed to be a proper AI system. It has expert knowledge of tax law and it applies that legal expertise to your specific situation. True, this is largely good old-fashioned AI already scientifically understood in the 1970s, but by its access to relevant databases of the interconnected-big-data kind, it certainly has a modern twist. One could even say that the system is grounded in real world data, and is hence an example of situated AI, in the way that the term was used in the 1990s (and perhaps before). But also this is clearly not an adaptive machine learning AI system, as is today expected of AI.

3 AI & law is hard

The reason why much of the successful legal tech is not really AI is simple. AI & Law is hard, very hard. In part this explains why many of us are here in this room. We are brave, we like the hard problems. In AI & Law they cannot be evaded.

Nederland ontwapent (The Netherlands disarm). Source: Nationaal Archief, 2.24.01.03, 918-0574 (Joost Evers, Anefo)

Let us look at an example of real law. We go back to the year when I was born when pacifism was still a relevant political attitude. In that year the Dutch Supreme court decided that the inscription ‘The Netherlands disarm’, mounted on a tower (Fig.  4 ) was not an offense. Footnote 16 The court admitted that indeed the sign could be considered a violation of Article 1 of the landscape management regulation of the province of North Holland, but the court decided that that regulation lacked binding power by a conflict with the freedom of speech, as codified in Article 7 of the Dutch constitution.

An example of a hard case. This outcome and its reasoning could not really be predicted, which is one reason why the case is still taught in law schools.

The example can be used to illustrate some of the tough hurdles for the development of AI & Law as they have been recognized from the start; here a list used by Rissland ( 1988 ) when reviewing Anne Gardner’s pioneering book ‘An AI approach to legal reasoning’ (Gardner 1987 ), a revision of her 1984 Stanford dissertation. Footnote 17 I am happy that both are present in this room today.

The subsumption model

Legal reasoning is rule-guided, rather than rule-governed In the example, indeed both the provincial regulation and the constitution were only guiding, not governing. Their conflict had to be resolved. A wise judge was needed.

Legal terms are open textured In the example it is quite a stretch to interpret a sign on a tower as an example of speech in the sense of freedom of speech, but that is what the court here did. It is the old puzzle of legally qualifying the facts, not at all an easy business, also not for humans. With my background in mathematics, I found legal qualification to be a surprisingly and unpleasantly underspecified problem when I took law school exams during my first years as assistant professor in legal informatics in Maastricht, back in the 1990s. Today computers also still would have a very hard time handling open texture.

Legal questions can have more than one answer, but a reasonable and timely answer must be given I have not checked how quickly the supreme court made its decision, probably not very quickly, but the case was settled. The conflict was resolved. A solution that had not yet been there, had been created, constructed. The decision changed a small part of the world.

The answers to legal questions can change over time In the example I am not sure about today’s law in this respect, in fact it is my guess that freedom of speech is still interpreted as broadly as here, and I would not be surprised when it is now interpreted even more broadly. But society definitely has changed since the late 1960s, and what I would be surprised about is when I would today see such a sign in the public environment.

One way of looking at the hurdles is by saying that the subsumption model is false. According to the subsumption model of law there is a set of laws, thought of as rules, there are some facts,—and you arrive at the legal answers, the legal consequences by applying the rules to the facts (Fig.  5 ). The case facts are subsumed under the rules, providing the legal solution to the case. It is often associated with Montesquieu’s phrase of the judge as a ‘bouche de la loi’, the mouth of the law, according to which a judge is just the one who makes the law speak.

All hurdles just mentioned show that this perspective cannot be true. Rules are only guiding, terms are open-textured, there can be more answers, and things can change.

The theory construction model (Verheij 2003a , 2005 )

Hence an alternative perspective on what happens when a case is decided. Legal decision making is a process of constructing and testing a theory, a series of hypotheses that are gradually developed and tested in a critical discussion (Fig.  6 ). The figure suggests an initial version of the facts, an initial version of the relevant rules, and an initial version of the legal conclusions. Gradually the initial hypothesis is adapted. Think of what happens in a court proceedings, and in what in the Netherlands is called the ‘raadkamer’, the internal discussion among judges, where after a careful constructive critical discussion—if the judges get the time for that of course—finally a tried and tested perspective on the case is arrived at, showing the final legal conclusions subsuming the final facts under the final rules. This is the picture I used in the 2003 AI & Law special issue of the AI journal, edited by Edwina Rissland, Kevin Ashley, and Ronald Loui, two of them here in this room. A later version with Floris Bex emphasises that also the perspective on the evidence and how it supports the facts is gradually constructed (Bex and Verheij 2012 ). In our field, the idea of theory construction in the law has for instance been emphasised by McCarty ( 1997 ), Hafner and Berman ( 2002 ), Gordon ( 1995 ), Bench-Capon and Sartor ( 2003 ) and Hage et al. ( 1993 ).

4 AI as law

Today’s claim is that good AI requires a different way of doing AI, a way that we in the field of AI & Law have been doing all along, namely doing AI in a way that meets the requirements of the law, in fact in a way that models how things are done in the law. Let us discuss this perspective a bit further.

There can be many metaphors on what AI is and how it should be done, as follows.

AI as mathematics, where the focus is on formal systems;

AI as technology, where the focus is on the art of system design;

AI as psychology, where the focus is on intelligent minds;

AI as sociology, where the focus is on societies of agents.

And then AI as law, to which we return in a minute (Table  1 ).

In AI as mathematics, one can think of the logical and probabilistic foundations of AI, indeed since the start and still now of core importance. It is said that the namegiver of the field of AI—John McCarty—thought of the foundations of AI as an instance of logic, and logic alone. In contrast today some consider AI to be a kind of statistics 2.0 or 3.0.

In AI as technology, one can think of meticulously crafted rule-based expert systems or of machine learning algorithms evaluated on large carefully labeled data sets. In AI as technology, AI applications and AI research meet most directly.

In AI as psychology, one can think of the modeling of human brains as in cognitive modeling, or of the smart human-like algorithms that are sometimes referred to as cognitive computing.

In AI as sociology, one can think of multi-agent systems simulating a society and of autonomous robots that fly in flocks.

Perhaps you have recognized the list of metaphors as the ones used by Toulmin ( 1958 ) when he discussed what he thought of as a crisis in the formal analysis of human reasoning. He argued that the classical formal logic then fashionable was too irrelevant for what reasoning actually was, and he arrived at a perspective of logic as law. Footnote 18 What he meant was that counterargument must be considered, that rules warranting argumentative steps are material—and not only formal—, that these rules are backed by factual circumstances, that conclusions are often qualified, uncertain, presumptive, and that reasoning and argument are to be thought of as the outcome of debates among individuals and in groups (see also Hitchcock and Verheij 2006 ; Verheij 2009 ). All of these ideas emphasised by Toulmin have now been studied extensively, with the field of AI & Law having played a significant role in the developments. Footnote 19

The metaphors can also be applied to the law, exposing some key ideas familiar in law.

If we think of law as mathematics, the focus is on the formality of procedural rule following and of stare decisis where things are well-defined and there is little room for freedom.

In law as technology, one can think of the art of doing law in a jurisdiction with either a focus on rules, as in civil law systems, or with a focus on cases, as in common law systems.

In law as psychology, one can think of the judicial reasoning by an individual judge, and of the judicial discretion that is to some extent allowed, even wanted.

In law as sociology, the role of critical discussion springs to mind, and of regulating a society in order to give order and prevent chaos.

And finally the somewhat pleonastic metaphor of law as law, but now as law in contrast with the other metaphors. I think of two specific and essential ideas in the law, namely that government is to be bound by the rule of law, and that the goal of law is to arrive at justice, thereby supporting a good society and a good life for its citizens.

Note how this discussion shows the typically legal, hybrid balancing of different sides: rules and cases, regulations and decisions, rationality and interpretation, individual and society, boundedness and justice. And as we know this balancing best takes place in a constructive critical discussion.

Which brings us to bottom line of the list of AI metaphors (Table  1 ).

5. AI as law, where the focus is on hybrid critical discussion.

In AI as law, AI systems are to be thought of as hybrid critical discussion systems, where different hypothetical perspectives are constructed and evaluated until a good answer is found.

Bridging the gap between knowledge and data systems in AI (Verheij 2018 )

In this connection, I recently explained what I think is needed in AI (Fig.  7 ), namely the much needed step we have to make towards hybrid systems that connect knowledge representation and reasoning techniques with the powers of machine learning. In this diagram I used the term argumentation systems. But since argumentation has a very specific sound in this community, and perhaps to some feels as a too specific, too limiting perspective, I today speak of AI as Law in the sense of the development of hybrid critical discussion systems.

5 Topics in AI

Let me continue with a discussion of core topics in AI with the AI as Law perspective in mind. My focus is on reasoning, knowledge, learning and language.

5.1 Reasoning

First reasoning. I then indeed think of argumentation where arguments and counterarguments meet (van Eemeren et al. 2014 ; Atkinson et al. 2017 ; Baroni et al. 2018 ). This is connected to the idea of defeasibility, where arguments become defeated when attacked by a stronger counterargument. Argumentation has been used to address the deep and old puzzles of inconsistency, incomplete information and uncertainty.

Here is an example argument about the Dutch bike owner Mary whose bike is stolen (Fig.  8 ). The bike is bought by John, hence both have a claim to ownership—Mary as the original owner, John as the buyer. But in this case the conflict can be resolved as John bought the bike for the low price of 20 euros, indicating that he was not a bona fide buyer. At such a price, he could have known that the bike was stolen, hence he has no claim to ownership as the buyer, and Mary is the owner.

• Argumentation

It is one achievement of the field of AI & Law that the logic of argumentation is by now well understood, so well that it can be implemented in argumentation diagramming software that applies the logic of argumentation, for instance the ArguMed software that I implemented long ago during my postdoc period in the Maastricht law school (Verheij 2003a , 2005 ). Footnote 20 It implements argumentation semantics of the stable kind in the sense of Dung’s abstract argumentation that was proposed some 25 years ago (Dung 1995 ), a turning point and a cornerstone in today’s understanding of argumentation, with many successes. Abstract argumentation also gave new puzzles such as the lack of standardization leading to all kinds of detailed comparative formal studies, and more fundamentally the multiple formal semantics puzzle. The stable, preferred, grounded and complete semantics were the four proposed by Dung ( 1995 ), quickly thereafter extended to 6 when the labeling-based stage and semi-stable semantics were proposed (Verheij 1996 ). But that was only the start because the field of computational argumentation was then still only emerging.

For me, it was obvious that a different approach was needed when I discovered that after combining attack and support 11 different semantics were formally possible (Verheij 2003b ), but practically almost all hardly relevant. No lawyer has to think about whether the applicable argumentation semantics is the semi-stable or the stage semantics.

One puzzle in the field is the following, here included after a discussion on the plane from Amsterdam to Montreal with Trevor Bench-Capon and Henry Prakken. A key idea underlying the original abstract argumentation paper is that derivation-like arguments can be abstracted from, allowing to focus only on attack. I know that for many this idea has helped them in their work and understanding of argumentation. For me, this was—from rather early on—more a distraction than an advantage as it introduced a separate, seemingly spurious layer. In the way that my PhD supervisor Jaap Hage put it: ‘those cloudy formal structures of yours’—and Jaap referred to abstract graphs in the sense of Dung—have no grounding in how lawyers think. There is no separate category of supporting arguments to be abstracted from before considering attack; instead, in the law there are only reasons for and against conclusions that must be balanced. Those were the days when Jaap Hage was working on Reason-Based Logic ( 1997 ) and I was helping him (Verheij et al. 1998 ). In a sense, the ArguMed software based on the DefLog formalism was my answer to removing that redundant intermediate layer (still present in its precursor the Argue! system), while sticking to the important mathematical analysis of reinstatement uncovered by Dung (see Verheij 2003a , 2005 ). For background on the puzzle of combining support and attack, see (van Eemeren et al. 2014 , Sect. 11.5.5).

But as I said from around the turn of the millenium I thought a new mathematical foundation was called for, and it took me years to arrive at something that really increased my understanding of argumentation: the case model formalism (Verheij 2017a , b ), but that is not for now.

5.2 Knowledge

The second topic of AI to be discussed is knowledge, so prominent in AI and in law. I then think of material, semi-formal argumentation schemes such as the witness testimony scheme, or the scheme for practical reasoning, as for instance collected in the nice volume by Walton et al. ( 2008 ).

I also think of norms, in our community often studied with a Hohfeldian or deontic logic perspective on rights and obligations as a background. Footnote 21 And then there are the ontologies that can capture large amounts of knowledge in a systematic way. Footnote 22

One lesson that I have taken home from working in the domain of law—and again don’t forget that I started in the field of mathematics where things are thought of as neat and clean—one lesson is that in the world of law things are always more complex than you think. One could say that it is the business of law to find the exactly right level of complexity, and that is often just a bit more complex than one’s initial idea. And if things are not yet complex now, they can become tomorrow. Remember the dynamics of theory construction that we saw earlier (Fig.  6 ).

Types of juristic facts (left); tree of individuals (right) (Hage and Verheij 1999 )

Figure  9 (left) shows how in the law different categories of juristic facts are distinguished. Here juristic facts are the kind of facts that are legally relevant, that have legal consequences. They come in two kinds: acts with legal consequences, and bare juristic facts, where the latter are intentionless events such as being born, which still have legal consequences. And acts with legal consequences are divided in on the one hand juristic acts aimed at a legal consequence (such as contracting), and on the other factual acts, where although there is no legal intention, still there are legal consequences. Here the primary example is that of unlawful acts as discussed in tort law. I am still happy that I learnt this categorization of juristic facts in the Maastricht law school, as it has relevantly expanded my understanding of how things work in the world. And of how things should be done in AI. Definitely not purely logically or purely statistically, definitely with much attention for the specifics of a situation.

Signing a sales contract (Hage and Verheij 1999 )

Figure  9 (right) shows another categorization, prepared with Jaap Hage, that shows how we then approached the core categories of things, or ‘individuals’ that should be distinguished when analyzing the law: states of affairs, events rules, other individuals, and then the subcategories of event occurrences, rule validities and other states of affairs. And although such a categorization does have a hint of the baroqueness of Jorge Luis Borges’ animal taxonomy (that included those animals that belong to the emperor, mermaids and innumerable animals), the abstract core ontology helped us to analyze the relations between events, rules and states of affairs that play a role when signing a contract (Fig.  10 ). Indeed at first sight a complex picture. For now it suffices that at the top row there is the physical act of signing—say when the pen is going over the paper to sign—and this physical act counts as engaging in a contractual bond (shown in the second row), which implies the undertaking of an obligation (third row), which in turn leads to a duty to perform an action (at the bottom row). Not a simple picture, but as said, in the law things are often more complex than expected, and typically for good, pragmatic reasons.

The core puzzle for our field and for AI generally that I would like to mention is that of commonsense knowledge. This remains an essential puzzle, also in these days of big data; also in these days of cognitive computing. Machines simply don’t have commonsense knowledge that is nearly good enough. A knowledgeable report in the Communications of the ACM explains that progress has been slow (Davis and Marcus 2015 ). It goes back to 2015, but please do not believe it when it is suggested that things are very different today. The commonsense knowledge problem remains a relevant and important research challenge indeed and I hope to see more of the big knowledge needed for serious AI & Law in the future. Only brave people have the chance to make real progress here, like the people in this room.

One example of what I think is an as yet underestimated cornerstone of commonsense knowledge is the role of globally coherent knowledge structures—such as the scenarios and cases we encounter in the law. Our current program chair Floris Bex took relevant steps to investigate scenario schemes and how they are hierarchically related, in the context of murder stories and crime investigation (Bex 2011 ). Footnote 23 Our field would benefit from more work like this, that goes back to the frames and scripts studied by people such as Roger Schank and Marvin Minsky.

My current favorite kind of knowledge representation uses the case models mentioned before. It has for instance been used to represent how an appellate court gradually constructs its hypotheses about a murder case on the basis of the evidence, gradually testing and selecting which scenario of what has happened to believe or not (Verheij 2019 ), and also to the temporal development of the relevance of past decisions in terms of the values they promote and demote (Verheij 2016 ).

5.3 Learning

Then we come to the topic of learning. It is the domain of statistical analysis that shows that certain judges are more prone to supporting democrat positions than others, and that as we saw no longer is allowed in France. It is the domain of open data, that allows public access to legal sources and in which our community has been very active (Biagioli et al. 2005 ; Francesconi and Passerini 2007 ; Francesconi et al. 2010a , b ; Sartor et al. 2011 ; Athan et al. 2013 ). And it is the realm of neural networks, back in the days called perceptrons, now referred to as deep learning.

The core theme to be discussed here is the issue of how learning and the justification of outcomes go together, using a contemporary term: how to arrive at an explainable AI, an explainable machine learning. We have heard it discussed at all career levels, by young PhD students and by a Turing award winner.

The issue can be illustrated by a mock prediction machine for Dutch criminal courts. Imagine a button that you can push, that once you push it always gives the outcome that the suspect is guilty as charged. And thinking of the need to evaluate systems (Conrad and Zeleznikow 2015 ), this system has indeed been validated by the Dutch Central Bureau of Statistics, that has the data that shows that this prediction machine is correct in 91 out of a 100 cases (Fig.  11 ). The validating data shows that the imaginary prediction machine has become a bit less accurate in recent years, presumably by changes in society, perhaps in part caused by the attention in the Netherlands for so-called dubious cases, or miscarriages of justice, which may have made judges a little more reluctant to decide for guilt. But still: 91% for this very simple machine is quite good. And as you know, all this says very little about how to decide for guilt or not.

Convictions in criminal cases in the Netherlands; source: Central Bureau of Statistics ( www.cbs.nl ), data collection of September 11, 2017

How hard judicial prediction really is, also when using serious machine learning techniques, is shown by some recent examples. Katz et al. ( 2017 ) that their US Supreme Court prediction machine could achieve a 70% accuracy. A mild improvement over the baseline of the historical majority outcome (to always affirm a previous decision) which is 60%, and even milder over the 10 year majority outcome which is 67%. The system based its predictions on features such as judge identity, month, court of origin and issue, so modest results are not surprising.

In another study Aletras and colleagues ( 2016 ) studied European Court of Human Rights cases. They used n-grams and topics as the starting point of their training, and used a prepared dataset to make a cleaner baseline of 50% accuracy by random guessing. They reached 79% accuracy using the whole text, and noted that by only using the part where the factual circumstances are described already an accuracy of 73% is reached.

Naively taking the ratios of 70 over 60 and of 79 over 50, one sees that factors of 1.2 and of 1.6 improvement are relevant research outcomes, but practically modest. And more importantly these systems only focus on outcome, without saying anything about how to arrive at an outcome, or about for which reasons an outcome is warranted or not.

And indeed and as said before learning is hard, especially in the domain of law. Footnote 24 I am still a fan of an old paper by Trevor Bench-Capon on neural networks and open texture (Bench-Capon 1993 ). In an artificially constructed example about welfare benefits, he included different kinds of constraints: boolean, categorical, numeric. For instance, women were allowed the benefit after 60, and men after 65. Trevor found that after training, the neural network could achieve a high overall performance, but with somewhat surprising underlying rationales. In Fig.  12 , on the left, one can see that the condition starts to be relevant long before the ages of 60 and 65 and that the difference in gender is something like 15 years instead of 5. On the right, with a more focused training set using cases with only single failing conditions, the relevance started a bit later, but still too early, while the gender difference now indeed was 5 years.

Neural networks and open texture (Bench-Capon 1993 )

What I have placed my bets on is the kind of hybrid cases and rules systems that for us in AI & Law are normal. Footnote 25 I now represent Dutch tort law in terms of case models validating rule-based arguments (Verheij 2017b ) (cf. Fig.  13 below).

5.4 Language

Then language, the fourth and final topic of AI that I would like to discuss with you. Today the topic of language is closely connected to machine learning. I think of the labeling of natural language data to allow for training; I think of prediction such as by a search engine or chat application on a smartphone, and I think of argument mining, a relevant topic with strong roots in the field of AI & Law.

The study of natural language in AI, and in fact of AI itself, got a significant boost by IBM’s Watson system that won the Jeopardy! quiz show. For instance, Watson correctly recognized the description of ‘A 2-word phrase [that] means the power to take private property for public use’. That description refers to the typically legal concept of eminent domain, the situation in which a government disowns property for public reasons, such as the construction of a highway or windmill park. Watson’s output showed that the legal concept scored 98%, but also ‘electric company’ and ‘capitalist economy’ were considered with 9% and 5% scores, respectively. Apparently Watson sees some kind of overlap between the legal concept of eminent domain, electric companies and capitalist economy, since 98+9+5 is more than a 100 percent.

And IBM continued, as Watson was used as the basis for its debating technologies. In a 2014 demonstration, Footnote 26 the system is considering the sale of violent video games to minors. The video shows that the system finds reasons for and against banning the sale of such games to minors, for instance that most children who play violent games do not have problems, but that violent video games can increase children’s aggression. The video remains impressive, and for the field of computational argumentation that I am a member of it was somewhat discomforting that the researchers behind this system were then outsiders to the field.

The success of these natural language systems leads one to think about why they can do what they do. Do they really have an understanding of a complex sentence describing the legal concept of eminent domain; can they really digest newspaper articles and other online resources on violent video games?

These questions are especially relevant since in our field of AI & Law we have had the opportunity to follow research on argument mining from the start. Early and relevant research is by Raquel Mochales Palau and Sien Moens, who studied argument mining in a paper at the 2009 ICAIL conference ( 2009 , 2011 ). And as already shown in that paper, it should not be considered an easy task to perform argument mining. Indeed the field has been making relevant and interesting progress, as also shown in research presented at this conference, but no one would claim the kind of natural language understanding needed for interpreting legal concepts or online debates. Footnote 27

So what then is the basis of apparent success? Is it simply because a big tech company can do a research investment that in academia one can only dream of? Certainly that is a part of what has been going on. But there is more to it than that as can be appreciated by a small experiment I did, this time actually an implemented online system. It is what I ironically called Poor Man’s Watson, Footnote 28 which has been programmed without much deep natural language technology, just some simple regular expression scripts using online access to the Google search engine and Wikipedia. And indeed it turns out that the simple script can also recognize the concept of eminent domain: when one types ‘the power to take private property for public use’ the answer is ‘eminent domain’. The explanation for this remarkable result is that for some descriptions the correct Wikipedia page ends up high in the list of pages returned by Google, and that happens because we—the people—have been typing in good descriptions of those concepts in Wikipedia, and indeed Google can find these pages. Sometimes the results are spectacular, but also they are brittle since seemingly small, irrelevant changes can quickly break this simple system.

And for the debating technology something similar holds since there are web sites collecting pros and cons of societal debates. For instance, the web site procon.org has a page on the pros and cons of violent video games. Footnote 29 Arguments it has collected include ‘Pro 1: Playing violent video games causes more aggression, bullying, and fighting’ and ‘Con 1: Sales of violent video games have significantly increased while violent juvenile crime rates have significantly decreased’. The web site Kialo has similar collaboratively created lists. Footnote 30 Concerning the issue ‘Violent video games should be banned to curb school shootings’, it lists for instance the pro ‘Video games normalize violence, especially in the eyes of kids, and affect how they see and interact with the world’ and the con ‘School shootings are, primarily, the result of other factors that should be dealt with instead’.

Surely the existence of such lists typed in, in a structured way, by humans is a central basis for what debating technology can and cannot do. It is not a coincidence that—listening carefully to the reports—the examples used in marketing concern curated lists of topics. At the same time this does not take away the bravery of IBM and how strongly it has been stimulating the field of AI by its successful demos. And that also for IBM things are sometimes hard is shown by the report from February 2019 when IBM’s technology entered into a debate with a human debater, and this time lost. Footnote 31 But who knows what the future brings.

What I believe is needed is the development of an ever closer connection between complex knowledge representations and natural language explanations, as for instance in work by Charlotte Vlek on explaining Bayesian Networks (Vlek et al. 2016 ), which had nice connections to the work discussed by Jeroen Keppens yesterday ( 2019 ).

6 Conclusion

As I said I think the way to go for the field is to develop an AI that is much like the law, an AI where systems are hybrid critical discussion systems.

For after phases of AI as mathematics, as technology, as psychology, and as sociology—all still important and relevant—, an AI as Law perspective provides fresh ideas for designing an AI that is good (Table  1 ). And in order to build the hybrid critical discussion systems that I think are needed, lots of work is waiting in reasoning, in knowledge, in learning and in language, as follows.

For reasoning (Sect.  5.1 ), the study of formal and computational argumentation remains relevant and promising, while work is needed to arrive at a formal semantics that is not only accessible for a small group of experts.

For knowledge (Sect.  5.2 ), we need to continue working on knowledge bases large and small, and on systems with embedded norms. But I hope that some of us are also brave enough to be looking for new ways to arrive at good commonsense knowledge for machines. In the law we cannot do without wise commonsense.

For learning (Sect.  5.3 ), the integration of knowledge and data can be addressed by how in the law rules and cases are connected and influence one another. Only then the requirements of explainability and responsibility can be properly addressed.

For language (Sect.  5.4 ), work is needed in interpretation of what is said in a text. This requires an understanding in terms of complex, detailed models of a situation, like what happens in any court of law where every word can make a relevant difference.

Lots of work to do. Lots of high mountains to conquer.

The perspective of AI as Law discussed here today can be regarded as an attempt to broaden what I said in the lecture on ‘Arguments for good AI’ where the focus is mostly on computational argumentation (Verheij 2018 ). There I explain that we need a good AI that can give good answers to our questions, give good reasons for them, and make good choices. I projected that in 2025 we will have arrived at a new kind of AI systems bridging knowledge and data, namely argumentation systems (Fig.  7 ). Clearly and as I tried to explain today, there is still plenty of work to be done. I expect that a key role will be played by work in our field on connections between rules, cases and arguments, as in the set of cases formalizing tort law (Fig.  13 , left) that formally validate the legally relevant rule-based arguments (Fig.  13 , right).

Arguments, rules and cases for Dutch tort law (Verheij 2017b )

By following the path of developing AI as Law we can guard against technology that is bad for us, and that—unlike the guillotine I started with—is a really humane technology that directly benefits society and its citizens.

In conclusion, in these days of dreams and fears of AI and algorithms, our beloved field of AI & Law is more relevant than ever. We can be proud that AI & Law has worked on the design of socially aware, explainable, responsible AI for decades already.

And since we in AI & Law are used to address the hardest problems across the breadth of AI (reasoning, knowledge, learning, language)—since in fact we cannot avoid them—, our field can inspire new solutions. In particular, I discussed computational argumentation, schemes for arguments and scenarios, encoded norms, hybrid rule-case systems and computational interpretation.

We only need to look at what happens in the law. In the law, we see an artificial system that adds much value to our life. Let us take inspiration from the law, and let us work on building Artificial Intelligence that is not scary, but that genuinely contributes to a good quality of life in a just society. I am happy and proud to be a member of this brave and smart community and I thank you for your attention.

This text is an adapted version of the IAAIL presidential address delivered at the 17th International Conference on Artificial Intelligence and Law (ICAIL 2019) in Montreal, Canada (Cyberjustice Lab, University of Montreal, June 19, 2019).

‘Beschonken Meppeler rijdt slapend over de snelweg’ (automatic translation: ‘Drunken Meppeler sleeps on the highway’), RTV Drenthe, May 17, 2019.

‘Can AI be a fair judge in court? Estonia thinks so’, Wired, March 25, 2019 (Eric Miller).

‘Het nieuwe datadieet van Google en Facebook’ (automatic translation: ‘The new data diet from Google and Facebook’, nrc.nl , May 11, 2019.

‘Zo stuurt en controleert China zijn burgers’ (automatic translation: ‘This is how China directs and controls its citizens’, nrc.nl , June 14, 2019.

bnvki.org/wp-content/uploads/2018/09/Dutch-AI-Manifesto.pdf .

claire-ai.org .

iaail.org .

wetten.overheid.nl .

www.om.nl/onderwerpen/boetebase .

uitspraken.rechtspraak.nl .

www.belastingdienst.nl .

‘ICT-project basisregistratie totaal mislukt’ (automatic translation: ‘IT project basic registration totally failed’), nrc.nl , July 17, 2017.

techindex.law.stanford.edu .

Supreme Court The Netherlands, January 24, 1967: Nederland ontwapent (The Netherlands disarm).

For more on the complexity of AI & Law, see for instance (Rissland 1983 ; Sergot et al. 1986 ; Bench-Capon et al. 1987 , 2012 ; Rissland and Ashley 1987 ; Oskamp et al. 1989 ; Ashley 1990 , 2017 ; van den Herik 1991 ; Berman and Hafner 1995 ; Loui and Norman 1995 ; Bench-Capon and Sartor 2003 ; Sartor 2005 ; Zurek and Araszkiewicz 2013 ; Lauritsen 2015 ).

Toulmin ( 1958 ) speaks of logic as mathematics, as technology, as psychology, as sociology and as law (jurisprudence).

See for instance the research by Prakken ( 1997 ), Sartor ( 2005 ), Gordon ( 1995 ), Bench-Capon ( 2003 ) and Atkinson and Bench-Capon ( 2006 ). Argumentation research in AI & Law is connected to the wider study of formal and computational argumentation, see for instance (Simari and Loui 1992 ; Pollock 1995 ; Vreeswijk 1997 ; Chesñevar et al. 2000 ). See also the handbooks (Baroni et al. 2018 ; van Eemeren et al. 2014 ).

For some other examples, see (Gordon et al. 2007 ; Loui et al. 1997 ; Kirschner et al. 2003 ; Reed and Rowe 2004 ; Scheuer et al. 2010 ; Lodder and Zelznikow 2005 ).

See for instance (Sartor 2005 ; Gabbay et al. 2013 ; Governatori and Rotolo 2010 ).

See for instance (McCarty 1989 ; Valente 1995 ; van Kralingen 1995 ; Visser 1995 ; Visser and Bench-Capon 1998 ; Hage and Verheij 1999 ; Boer et al. 2002 , 2003 ; Breuker et al. 2004 ; Hoekstra et al. 2007 ; Wyner 2008 ; Casanovas et al. 2016 ).

For more work on evidence in AI & Law, see for instance (Keppens and Schafer 2006 ; Bex et al. 2010 ; Keppens 2012 ; Fenton et al. 2013 ; Vlek et al. 2014 ; Di Bello and Verheij 2018 ).

See also recently (Medvedeva et al. 2019 ).

See for instance work by (Branting 1991 ; Skalak and Rissland 1992 ; Branting 1993 ; Prakken and Sartor 1996 , 1998 ; Stranieri et al. 1999 ; Roth 2003 ; Brüninghaus and Ashley 2003 ; Atkinson and Bench-Capon 2006 ; Čyras et al. 2016 ).

Milken Institute Global Conference 2014, session ‘Why Tomorrow Won’t Look Like Today: Things that Will Blow Your Mind’, youtu.be/6fJOtAzICzw?t=2725 .

See for instance (Schweighofer et al. 2001 ; Wyner et al. 2009 , 2010 ; Grabmair and Ashley 2011 ; Ashley and Walker 2013 ; Grabmair et al. 2015 ; Tran et al. 2020 ).

Poor Man’s Watson, www.ai.rug.nl/~verheij/pmw .

videogames.procon.org .

kialo.com .

‘IBM’s AI loses debate to a human, but it’s got worlds to conquer’, cnet.com , February 11, 2019.

Aletras N, Tsarapatsanis D, Preoţiuc-Pietro D, Lampos V (2016) Predicting judicial decisions of the European Court of Human Rights: a natural language processing perspective. Peer J Comput Sci 2:1–19. https://doi.org/10.7717/peerj-cs.93

Article   Google Scholar  

Ashley KD (1990) Modeling legal arguments: reasoning with cases and hypotheticals. The MIT Press, Cambridge

Google Scholar  

Ashley KD (2017) Artificial intelligence and legal analytics: new tools for law practice in the digital age. Cambridge University Press, Cambridge

Book   Google Scholar  

Ashley KD, Walker VR (2013) Toward constructing evidence-based legal arguments using legal decision documents and machine learning. In: Proceedings of the fourteenth international conference on artificial intelligence and law, pp 176–180. ACM, New York (New York)

Athan T, Boley H, Governatori G, Palmirani M, Paschke A, Wyner A (2013) OASIS LegalRuleML. In: Proceedings of the 14th international conference on artificial intelligence and law (ICAIL 2013), pp 3–12. ACM Press, New York (New York)

Atkinson K, Bench-Capon TJM (2006) Legal case-based reasoning as practical reasoning. Artif Intell Law 13:93–131

Article   MATH   Google Scholar  

Atkinson K, Baroni P, Giacomin M, Hunter A, Prakken H, Reed C, Simari G, Thimm M, Villata S (2017) Toward artificial argumentation. AI Mag 38(3):25–36

Baroni P, Gabbay D, Giacomin M, van der Torre L (eds) (2018) Handbook of formal argumentation. College Publications, London

MATH   Google Scholar  

Bench-Capon TJM (1993) Neural networks and open texture. In: Proceedings of the fourth international conference on artificial intelligence and law, pp 292–297. ACM Press, New York (New York)

Bench-Capon TJM (2003) Persuasion in practical argument using value-based argumentation frameworks. J Logic Comput 13(3):429–448

Article   MathSciNet   MATH   Google Scholar  

Bench-Capon TJM, Sartor G (2003) A model of legal reasoning with cases incorporating theories and values. Artif Intell 150(1):97–143

Bench-Capon TJM, Robinson GO, Routen TW, Sergot MJ (1987) Logic programming for large scale applications in law: a formalisation of supplementary benefit legislation. In: Proceedings of the 1st international conference on artificial intelligence and law (ICAIL 1987), pp 190–198. ACM, New York (New York)

Bench-Capon T, Araszkiewicz M, Ashley KD, Atkinson K, Bex FJ, Borges F, Bourcier D, Bourgine D, Conrad JG, Francesconi E, Gordon TF, Governatori G, Leidner JL, Lewis DD, Loui RP, McCarty LT, Prakken H, Schilder F, Schweighofer E, Thompson P, Tyrrell A, Verheij B, Walton DN, Wyner AZ (2012) A history of AI and Law in 50 papers: 25 years of the international conference on AI and law. Artif Intell Law 20(3):215–319

Berman DH, Hafner CL (1995) Understanding precedents in a temporal context of evolving legal doctrine. In: Proceedings of the fifth international conference on artificial intelligence and law, pp 42–51. ACM Press, New York (New York)

Bex FJ (2011) Arguments, stories and criminal evidence: a formal hybrid theory. Springer, Berlin

Bex FJ, Verheij B (2012) Solving a murder case by asking critical questions: an approach to fact-finding in terms of argumentation and story schemes. Argumentation 26:325–353

Bex FJ, van Koppen PJ, Prakken H, Verheij B (2010) A hybrid formal theory of arguments, stories and criminal evidence. Artif Intell Law 18:1–30

Biagioli C, Francesconi E, Passerini A, Montemagni S, Soria C (2005) Automatic semantics extraction in law documents. In: Proceedings of the 10th international conference on artificial intelligence and law (ICAIL 2005), pp 133–140. ACM Press, New York (New York)

Boer A, Hoekstra R, Winkels R (2002) METAlex: legislation in XML. In: Bench-Capon TJM, Daskalopulu A, Winkels R (eds) Legal knowledge and information systems. JURIX 2002: the fifteenth annual conference. IOS Press, Amsterdam, pp 1–10

Boer A, van Engers T, Winkels R (2003) Using ontologies for comparing and harmonizing legislation. In: Proceedings of the 9th international conference on artificial intelligence and law, pp 60–69. ACM, New York (New York)

Branting LK (1991) Building explanations from rules and structured cases. Int J Man Mach Stud 34(6):797–837

Branting LK (1993) A computational model of ratio decidendi. Artif Intell Law 2(1):1–31

Breuker J, Valente A, Winkels R (2004) Legal ontologies in knowledge engineering and information management. Artif Intell Law 12(4):241–277

Brüninghaus S, Ashley KD (2003) Predicting outcomes of case based legal arguments. In: Proceedings of the 9th international conference on artificial intelligence and law (ICAIL 2003), pp 233–242. ACM, New York (New York)

Casanovas P, Palmirani M, Peroni S, van Engers T, Vitali F (2016) Semantic web for the legal domain: the next step. Semant Web 7(3):213–227

Chesñevar CI, Maguitman AG, Loui RP (2000) Logical models of argument. ACM Comput Surv 32(4):337–383

Conrad JG, Zeleznikow J (2015) The role of evaluation in ai and law: an examination of its different forms in the ai and law journal. In: Proceedings of the 15th international conference on artificial intelligence and law (ICAIL 2015), pp 181–186. ACM, New York (New York)

Čyras K, Satoh K, Toni F (2016) Abstract argumentation for case-based reasoning. In: Proceedings of the fifteenth international conference on principles of knowledge representation and reasoning (KR 2016), pp 549–552. AAAI Press

Davis E, Marcus G (2015) Commonsense reasoning and commonsense knowledge in artificial intelligence. Commun ACM 58(9):92–103

Di Bello M, Verheij B (2018) Evidential reasoning. In: Bongiovanni G, Postema G, Rotolo A, Sartor G, Valentini C, Walton DN (eds) Handbook of legal reasoning and argumentation. Springer, Dordrecht, pp 447–493

Chapter   Google Scholar  

Dung PM (1995) On the acceptability of arguments and its fundamental role in nonmonotonic reasoning, logic programming and n-person games. Artif Intell 77:321–357

Fenton NE, Neil MD, Lagnado DA (2013) A general structure for legal arguments about evidence using Bayesian Networks. Cognit Sci 37:61–102

Francesconi E, Passerini A (2007) Automatic classification of provisions in legislative texts. Artif Intell Law 15(1):1–17

Francesconi E, Montemagni S, Peters W, Tiscornia D (2010a) Integrating a bottom–up and top–down methodology for building semantic resources for the multilingual legal domain. In: Semantic processing of legal texts, pp 95–121. Springer, Berlin

Francesconi E, Montemagni S, Peters W, Tiscornia D (2010b) Semantic processing of legal texts: where the language of law meets the law of language. Springer, Berlin

Gabbay D, Horty J, Parent X, Van der Meyden R, van der Torre L (2013) Handbook of deontic logic and normative systems. College Publication, London

Gardner A (1987) An artificial intelligence approach to legal reasoning. The MIT Press, Cambridge

Gordon TF (1995) The pleadings game: an artificial intelligence model of procedural justice. Kluwer, Dordrecht

Gordon TF, Prakken H, Walton DN (2007) The Carneades model of argument and burden of proof. Artif Intell 171(10–15):875–896

Governatori G, Rotolo A (2010) Changing legal systems: legal abrogations and annulments in defeasible logic. Logic J IGPL 18(1):157–194

Grabmair M, Ashley KD (2011) Facilitating case comparison using value judgments and intermediate legal concepts. In: Proceedings of the 13th international conference on Artificial intelligence and law, pp 161–170. ACM, New York (New York)

Grabmair M, Ashley KD, Chen R, Sureshkumar P, Wang C, Nyberg E, Walker VR (2015) Introducing LUIMA: an experiment in legal conceptual retrieval of vaccine injury decisions using a uima type system and tools. In: Proceedings of the 15th international conference on artificial intelligence and law, pp 69–78. ACM, New York (New York)

Hafner CL, Berman DH (2002) The role of context in case-based legal reasoning: teleological, temporal, and procedural. Artif Intell Law 10(1–3):19–64

Hage JC (1997) Reasoning with rules. An essay on legal reasoning and its underlying logic. Kluwer Academic Publishers, Dordrecht

Hage JC, Verheij B (1999) The law as a dynamic interconnected system of states of affairs: a legal top ontology. Int J Hum Comput Stud 51(6):1043–1077

Hage JC, Leenes R, Lodder AR (1993) Hard cases: a procedural approach. Artif Intell Law 2(2):113–167

Hitchcock DL, Verheij B (eds) (2006) Arguing on the toulmin model. New essays in argument analysis and evaluation (argumentation library, volume 10). Springer, Dordrecht

Hoekstra R, Breuker J, Di Bello M, Boer A (2007) The lkif core ontology of basic legal concepts. In: Casanovas P, Biasiotti MA, Francesconi E, Sagri MT (eds). Proceedings of LOAIT 2007. Second workshop on legal ontologies and artificial intelligence techniques, pp 43–63. CEUR-WS

Katz DM, Bommarito II MJ, Blackman J (2017) A general approach for predicting the behavior of the Supreme Court of the United States. PLoS ONE 12(4):1–18. https://doi.org/10.1371/journal.pone.0174698

Keppens J (2012) Argument diagram extraction from evidential Bayesian networks. Artif Intell Law 20:109–143

Keppens J (2019) Explainable Bayesian network query results via natural language generation systems. In: Proceedings of the 17th international conference on artificial intelligence and law (ICAIL 2019), pp 42–51. ACM, New York (New York)

Keppens J, Schafer B (2006) Knowledge based crime scenario modelling. Expert Syst Appl 30(2):203–222

Kirschner PA, Shum SJB, Carr CS (2003) Visualizing argumentation: software tools for collaborative and educational sense-making. Springer, Berlin

Lauritsen M (2015) On balance. Artif Intell Law 23(1):23–42

Lodder AR, Zelznikow J (2005) Developing an online dispute resolution environment: dialogue tools and negotiation support systems in a three-step model. Harvard Negot Law Rev 10:287–337

Loui RP, Norman J (1995) Rationales and argument moves. Artif Intell Law 3:159–189

Loui RP, Norman J, Altepeter J, Pinkard D, Craven D, Linsday J, Foltz M (1997) Progress on room 5: a testbed for public interactive semi-formal legal argumentation. In: Proceedings of the 6th international conference on artificial intelligence and law, pp 207–214. ACM Press

McCarty LT (1989) A language for legal discourse. i. basic features. In: Proceedings of the 2nd international conference on artificial intelligence and law (ICAIL 1989), pp 180–189. ACM, New York (New York)

McCarty LT (1997) Some arguments about legal arguments. In: Proceedings of the 6th international conference on artificial intelligence and law (ICAIL 1997), pp 215–224. ACM Press, New York (New York)

Medvedeva M, Vols M, Wieling M (2019) Using machine learning to predict decisions of the European court of human rights. Artif Intell Law. https://doi.org/10.1007/s10506-019-09255-y

Mochales Palau R, Moens MF (2009) Argumentation mining: the detection, classification and structure of arguments in text. In: Proceedings of the 12th international conference on artificial intelligence and law (ICAIL 2009), pp ges 98–107. ACM Press, New York (New York)

Mochales Palau R, Moens MF (2011) Argumentation mining. Artif Intell Law 19(1):1–22

Oskamp A, Walker RF, Schrickx JA, van den Berg PH (1989) PROLEXS divide and rule: a legal application. In: Proceedings of the second international conference on artificial intelligence and law, pp 54–62. ACM, New York (New York)

Pollock JL (1995) Cognitive carpentry: a blueprint for how to build a person. The MIT Press, Cambridge

Prakken H (1997) Logical tools for modelling legal argument. A study of defeasible reasonong in law. Kluwer Academic Publishers, Dordrecht

Prakken H, Sartor G (1996) A dialectical model of assessing conflicting arguments in legal reasoning. Artif Intell Law 4:331–368

Prakken H, Sartor G (1998) Modelling reasoning with precedents in a formal dialogue game. Artif Intell Law 6:231–287

Reed C, Rowe G (2004) Araucaria: software for argument analysis, diagramming and representation. Int J AI Tools 14(3–4):961–980

Rissland EL (1983) Examples in legal reasoning: Legal hypotheticals. In: Proceedings of the 8th international joint conference on artificial intelligence (IJCAI 1983), pp 90–93

Rissland EL (1988) Book review. An artificial intelligence approach to legal reasoning. Harvard J Law Technol 1(Spring):223–231

Rissland EL, Ashley KD (1987) A case-based system for trade secrets law. In: Proceedings of the first international conference on artificial intelligence and law, pp 60–66. ACM Press, New York (New York)

Roth B (2003) Case-based reasoning in the law. A formal theory of reasoning by case comparison. Dissertation Universiteit Maastricht, Maastricht

Sartor G (2005) Legal reasoning: a cognitive approach to the law. Vol 5 of Treatise on legal philosophy and general jurisprudence. Springer, Berlin

Sartor G, Palmirani M, Francesconi E, Biasiotti MA (2011) Legislative XML for the semantic web: principles, models, standards for document management. Springer, Berlin

Scheuer O, Loll F, Pinkwart N, McLaren BM (2010) Computer-supported argumentation: a review of the state of the art. Int J Comput Support Collab Learn 5(1):43–102

Schweighofer E, Rauber A, Dittenbach M (2001) Automatic text representation, classification and labeling in European law. In: Proceedings of the 8th international conference on artificial intelligence and law, pp 78–87. ACM, New York (New York)

Sergot MJ, Sadri F, Kowalski RA, Kriwaczek F, Hammond P, Cory HT (1986) The british nationality act as a logic program. Commun ACM 29(5):370–386

Simari GR, Loui RP (1992) A mathematical treatment of defeasible reasoning and its applications. Artif Intell 53:125–157

Skalak DB, Rissland EL (1992) Arguments and cases: an inevitable intertwining. Artif Intell Law 1(1):3–44

Stranieri A, Zeleznikow J, Gawler M, Lewis B (1999) A hybrid rule-neural approach for the automation of legal reasoning in the discretionary domain of family law in australia. Artif Intell Law 7(2–3):153–183

Toulmin SE (1958) The uses of argument. Cambridge University Press, Cambridge

Tran V, Le Nguyen M, Tojo S, Satoh K (2020) Encoded summarization: summarizing documents into continuous vector space for legal case retrieval. Artif Intell Law. https://doi.org/10.1007/s10506-020-09262-4

Valente A (1995) Legal knowledge engineering. A modelling approach. IOS Press, Amsterdam

van den Herik HJ (1991) Kunnen computers rechtspreken?. Gouda Quint, Arnhem

van Eemeren FH, Garssen B, Krabbe ECW, Snoeck Henkemans AF, Verheij B, Wagemans JHM (2014) Handbook of argumentation theory. Springer, Berlin

van Kralingen RW (1995) Frame-based conceptual models of statute law. Kluwer Law International, The Hague

Verheij B (1996) Two approaches to dialectical argumentation: admissible sets and argumentation stages. In: Meyer JJ, van der Gaag LC (eds) Proceedings of NAIC’96. Universiteit Utrecht, Utrecht, pp 357–368

Verheij B (2003a) Artificial argument assistants for defeasible argumentation. Artif Intell 150(1–2):291–324

Verheij B (2003b) DefLog: on the logical interpretation of prima facie justified assumptions. J Logic Comput 13(3):319–346

Verheij B (2005) Virtual arguments. On the design of argument assistants for lawyers and other arguers. T.M.C. Asser Press, The Hague

Verheij B (2009) The Toulmin argument model in artificial intelligence. Or: how semi-formal, defeasible argumentation schemes creep into logic. In: Rahwan I, Simari GR (eds) Argumentation in artificial intelligence. Springer, Berlin, pp 219–238

Verheij B (2016) Formalizing value-guided argumentation for ethical systems design. Artif Intell Law 24(4):387–407

Verheij B (2017a) Proof with and without probabilities. Correct evidential reasoning with presumptive arguments, coherent hypotheses and degrees of uncertainty. Artif Intell Law 25(1):127–154

Verheij B (2017b) Formalizing arguments, rules and cases. In: Proceedings of the 16th international conference on artificial intelligence and law (ICAIL 2017), pp 199–208. ACM Press, New York (New York)

Verheij B (2018) Arguments for good artificial intelligence. University of Groningen, Groningen. http://www.ai.rug.nl/~verheij/oratie/

Verheij B (2019) Analyzing the Simonshaven case with and without probabilities. Top Cognit Sci. https://doi.org/10.1111/tops.12436

Verheij B, Hage JC, van den Herik HJ (1998) An integrated view on rules and principles. Artif Intell Law 6(1):3–26

Visser PRS (1995) Knowledge specification for multiple legal tasks; a case study of the interaction problem in the legal domain. Kluwer Law International, The Hague

Visser PRS, Bench-Capon TJM (1998) A comparison of four ontologies for the design of legal knowledge systems. Artif Intell Law 6(1):27–57

Vlek CS, Prakken H, Renooij S, Verheij B (2014) Building Bayesian Networks for legal evidence with narratives: a case study evaluation. Artif Intell Law 22(4):375–421

Vlek CS, Prakken H, Renooij S, Verheij B (2016) A method for explaining Bayesian Networks for legal evidence with scenarios. Artif Intell Law 24(3):285–324

Vreeswijk GAW (1997) Abstract argumentation systems. Artif Intell 90:225–279

Walton DN, Reed C, Macagno F (2008) Argumentation schemes. Cambridge University Press, Cambridge

Book   MATH   Google Scholar  

Wyner A (2008) An ontology in OWL for legal case-based reasoning. Artif Intell Law 16(4):361

Wyner A, Angelov K, Barzdins G, Damljanovic D, Davis B, Fuchs N, Hoefler S, Jones K, Kaljurand K, Kuhn T et al (2009) On controlled natural languages: properties and prospects. In: International workshop on controlled natural language, pp 281–289. Berlin

Wyner A, Mochales-Palau R, Moens MF, Milward D (2010) Approaches to text mining arguments from legal cases. In: Semantic processing of legal texts, pp 60–79. Springer, Berlin

Zurek T, Araszkiewicz M (2013) Modeling teleological interpretation. In: Proceedings of the fourteenth international conference on artificial intelligence and law, pp 160–168. ACM, New York (New York)

Download references

Author information

Authors and affiliations.

Department of Artificial Intelligence, Bernoulli Institute of Mathematics, Computer Science and Artificial Intelligence, University of Groningen, Groningen, The Netherlands

Bart Verheij

You can also search for this author in PubMed   Google Scholar

Corresponding author

Correspondence to Bart Verheij .

Additional information

Publisher's note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ .

Reprints and Permissions

About this article

Verheij, B. Artificial intelligence as law. Artif Intell Law 28 , 181–206 (2020). https://doi.org/10.1007/s10506-020-09266-0

Download citation

Published : 14 May 2020

Issue Date : June 2020

DOI : https://doi.org/10.1007/s10506-020-09266-0

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Artificial intelligence
• Knowledge representation and reasoning
• Machine learning
• Natural language processing
• Find a journal
• Publish with us

Artificial Intelligence: Ethical, Social, Legal Issues

Ethical issues, social issues, professional and legal issues, reference list.

Artificial intelligence (AI) is an exciting but very controversial field of informational technology. Some sources say that the further development of this field will be useful for the humankind and will help us to solve many problems, for example, find cures for more diseases, increase the lifespan, give new possibilities in the space travel and so on. At the same time, a lot of other sources claim that such kind of technologies will be harmful to their own creators and that the only good superintelligent machine is the unplugged one. The field of artificial intelligence indeed brings numerous ethical, social, professional and legal issues; but are those so disturbing as some people claim?

Let us start with ethical issues. The greatest concern is this regard is the threat to security. Any AI program, regardless of the level of intelligence it demonstrates, remains only a software. Thus, it has all the drawbacks that the software has. First of all, an AI program can be copied – as long as there are people who can do this and the hardware that can store it (Bostrom 2003). Surely, it can be not easy or quick but it can happen. And that is how valuable data can get into the wrong hands. Secondly, machines can make mistakes, which threats the security as well.

As an example, let us imagine that there is an intelligent vision program that scans people’s baggage at the airport (Bostrom & Yudkowsky 2014, p. 317). What if there is the flaw in the algorithms, because of which the program is unable to recognize a bomb if a pistol is put next to it? Such kind of a mistake is possible, and it threatens the security and safety of every person on the board.

Another great concern includes privacy issues. The boundary between the operation of AI programs and the violation of people’s privacy is rather thin. There are a lot of examples when artificial intelligence has been involved in large disputes because of the privacy violations. As Weeks (2012) writes in The Globe and Mail , many businesses use AI programs to collect and store data about their customers: personal data taken from social networks, the location information, shopping patterns, payment habits, and so on. They try to ‘track their customers’ to increase the profit but for a regular user that usually means the violation of his or her privacy (Weeks 2012, para. 3).

Joel Rosenblatt (2014) describes another example of privacy violation for marketing purposes: aiming to advertise their services to a greater number of customers, LinkedIn Corp. downloaded the contacts from their customers’ external e-mails used as usernames on the LinkedIn site and sent several advertising letters to those email addresses. Apart from AI applications in advertising, many other programs are argued to violate users’ privacy. A prime example is the speech recognition technology. The latest smartphones are able to recognize their users by voice, which, as some people claim, steals users’ identity: ‘Your voice doesn’t just give away who you are, but what you’re like and what you’re doing … Your speech is like your fingerprints or your DNA’ (Rutkin 2015, para. 7).

Finally, the development of intelligent technologies can gradually lead to the development of even more intelligent ones, so-called advanced intelligence or superintelligence, and that can happen suddenly (Bostrom 2003). Although it will probably solve or at least help the humankind to solve many problems, including poverty, incurable diseases, global environmental problems, and so on, if it is used for evil purposes, it can exacerbate many other problems.

As a prime example, AI can contribute to wars by creating advanced weaponry – autonomous weapons and military robots (Romportl, Zackova & Kelemen 2014). Some semi-autonomous weapons have already been used by the United States and North Korea, for example (Romportl, Zackova & Kelemen 2014). But even though these weapons did some part of work by themselves (identified the target, for instance), they were not fully autonomous.

With all of this in mind, the question arises: do we actually need artificial intelligence or is it safer to delay its development? The truth is there is also the other side of the debate. While many people claim that AI technologies constitute a threat to security, others prove that they can help to strengthen it as well. In his article for the 3rd International Conference on Cyber Conflict, Enn Tyugu (2011) explains why AI technologies are one of the best options to defend the cyberspace.

Considering the speed of the processes in the cyber defense, as well as the amount of data transferred, without at least the minimum atomization, people are unable to provide an appropriate defense. The situation becomes even more difficult in view of the development and intelligence of modern malware and the frequency and sophistication of cyber-attacks (Tyugu 2011, p. 102). To handle all of this and be able to fight back, people need intelligent defense methods; otherwise, forces are not equal.

Additionally, many ethical issues and risks associated with artificial intelligence can be eliminated or at least minimized by particular precautions. Bostrom and Yudkowsky (2014) write that, in order to be safe, AI technologies should be predictable and transparent to inspections (to make the error detection possible and simple) and robust to manipulations to avoid such situations as the one with a bomb and a pistol described above. Although artificial intelligence is fraught with some new ethical problems and concerns, those can and should be solved as it has already been done with any other technological development.

Apart from ethical issues, the literature also talks about social ones. One of the most important is that artificial intelligence should contribute to peace, harmony, the wellbeing and the development of the population rather than increase the number of wars and battles in the world. AI applications and techniques improve medicine through analyzing complex data, helping with the diagnosis, treatment and even the prediction of the patient outcomes (Ramesh et al. 2004). Moreover, those can be used in almost any field of medicine (Ramesh et al. 2004). Artificial intelligence improves education by automating basic activities, providing tutoring AI programs, finding gaps in courses, and so on ( 10 Roles For Artificial Intelligence In Education 2015).

It can even contribute to psychology and help with people’s relations. In their article, Hergovich and Olbrich (2002) provide the review of several AI technologies that help to predict conflicts and calculate probabilities of those, determine the course of a dispute, and create peace negotiations for arguments. Finally, AI greatly accelerates the development of science. On the other hand, it can also be used to create advanced weapons. And the higher level of development this field of science reaches, the more impact on our society it has.

Another significant social concern is how people will interact with superintelligent machines if those are discovered. Kizza (2013) writes about a social paradox: people want to create machines that will do their work, but they do not want these machines to become too good in this work (p. 206). If they do achieve better intelligence than humans have, people become afraid of them, which brings many questions about the cooperation between people and intelligent agents created by them.

However, Ramos, Augusto, and Shapiro (2008) state that this problem can be addressed with the help of ambient intelligence, which helps to create technologies that are sensitive and responsive to people’s presence. If superintelligent machines are aware of people’s needs, can adapt to different environments (such as houses, schools, hospitals, offices, sports, and so on) and interact with people, they will have more chances to be accepted by their creators (Ramos, Augusto & Shapiro, 2008).

Finally, the last controversial question discussed in the literature is the one connected to professional and legal aspects. As Elkins (2015) writes in her article in Business Insider , experts predict that by 2025 robots will take over one-third of all people’s professions. Even now, robots work in health care centers assisting doctors and start to learn some white-collar jobs, which earlier seemed to be challenging for them (Elkins 2015). If the predicted outcome happens, and one-third of people’s jobs are performed by intelligent machines, it will not only change our economic system significantly but will also bring many legal issues. First of all, should robots be given the same rights as people have? Should they be protected by the Constitution or provided with full civil rights, which includes the right to reproduce (Elkins 2015)?

Additionally, if an intelligent machine commits a crime, should it be responsible for it, as a human being is responsible for theirs? Although it sounds unrealistic and even funny, something like that has already happened. An intelligent shopping robot developed for purchasing purposes managed to buy a Hungarian passport and Ecstasy pills (Elkins 2015). That time, a robot has not been charged (Elkins 2015). However, if something like this happens again, who is responsible for that? While many people claim that intelligent agents should be given with more or less the same rights and responsibilities that people have, Jack Millner (2015) believes, and I agree, that robots will need new legislations established for them.

To conclude, the development of artificial intelligence is indeed fraught with many controversial questions and problems. AI can give our society many positive things such as the advanced education, medicine, breakthroughs in science, and so on. At the same time, it increases the risk of wars and brings numerous unsolved professional and legal issues. Nevertheless, from my point of view, the greatest problem about artificial intelligence is the humankind, even though it sounds paradoxical. People have always been afraid of everything new, and the possibility of the creation of machines, which are more intelligent than humans, is frightening as well. Besides, many ethical and social problems can be solved, and even legal issues can be regulated. The only really insoluble problem is the one about wars and weaponry. That is highly unlikely that superintelligent technologies will take over the world and destroy it, but people can do it by themselves.

10 Roles For Artificial Intelligence In Education 2015.

Bostrom, N & Yudkowsky, E 2014, ‘The ethics of artificial intelligence’, in K Frankish & W Ramsey (eds), The Cambridge Handbook of Artificial Intelligence , Cambridge University Press, Cambridge, pp. 316-334.

Bostrom, N 2003, ‘Ethical Issues in Advanced Artificial Intelligence’, Cognitive, Emotive and Ethical Aspects of Decision Making in Humans and in Artificial Intelligence , vol. 2, no. 1, pp. 12-17.

Elkins, K 2015, Experts predict robots will take over 30% of our jobs by 2025 — and white-collar jobs aren’t immune .

Hergovich, A & Olbrich, A 2002, ‘What can Artificial Intelligence do for Peace Psychology?’, Review of Psychology , vol. 9, no. 1-2, pp. 3-11.

Kizza, JM 2013, Ethical and Social Issues in the Information Age , 5th edn, Springer, New York, New York.

Millner, J 2015, Should robots have human rights?

Ramesh, AN, Kambhampati, C, Monson, JR & Drew, PJ 2004, ‘Artificial intelligence in medicine’, Annals of the Royal College of Surgeons of England , vol. 86, no. 5, pp. 334-338.

Ramos, C, Augusto, JC & Shapiro, D 2008, ‘Ambient Intelligence – the Next Step for Artificial Intelligence’, IEEE Computer Society , vol. 23, no. 2, pp. 15-18.

Romportl, J, Zackova, E & Kelemen, J 2014, Beyond Artificial Intelligence: The Disappearing Human-Machine Divide , Springer, New York, New York.

Rosenblatt, J 2014, LinkedIn Ordered to Face Customer E-Mail Contacts Lawsuit. Web.

Rutkin, A 2015, Speech recognition AI identifies you by voice wherever you are .

Tyugu, E 2011, Artificial Intelligence in Cyber Defense , Web.

Weeks, C 2012, Dear valued customer, thank you for giving us all of your personal data .

Cite this paper

• Chicago (N-B)
• Chicago (A-D)

StudyCorgi. (2022, October 11). Artificial Intelligence: Ethical, Social, Legal Issues. Retrieved from https://studycorgi.com/artificial-intelligence-ethical-social-legal-issues/

StudyCorgi. (2022, October 11). Artificial Intelligence: Ethical, Social, Legal Issues. https://studycorgi.com/artificial-intelligence-ethical-social-legal-issues/

"Artificial Intelligence: Ethical, Social, Legal Issues." StudyCorgi , 11 Oct. 2022, studycorgi.com/artificial-intelligence-ethical-social-legal-issues/.

1. StudyCorgi . "Artificial Intelligence: Ethical, Social, Legal Issues." October 11, 2022. https://studycorgi.com/artificial-intelligence-ethical-social-legal-issues/.

Bibliography

StudyCorgi . "Artificial Intelligence: Ethical, Social, Legal Issues." October 11, 2022. https://studycorgi.com/artificial-intelligence-ethical-social-legal-issues/.

StudyCorgi . 2022. "Artificial Intelligence: Ethical, Social, Legal Issues." October 11, 2022. https://studycorgi.com/artificial-intelligence-ethical-social-legal-issues/.

StudyCorgi . (2022) 'Artificial Intelligence: Ethical, Social, Legal Issues'. 11 October.

This paper, “Artificial Intelligence: Ethical, Social, Legal Issues”, was written and voluntary submitted to our free essay database by a straight-A student. Please ensure you properly reference the paper if you're using it to write your assignment.

Before publication, the StudyCorgi editorial team proofread and checked the paper to make sure it meets the highest standards in terms of grammar, punctuation, style, fact accuracy, copyright issues, and inclusive language.

If you are the author of this paper and no longer wish to have it published on StudyCorgi, request the removal . Please use the “ Donate your paper ” form to submit an essay.