Insights

Integrated Artificial Intelligence in Law and Potential Risks & Uncertainties

Introduction

In the first place, Artificial Intelligence (“AI”) is a novel technical science that simulates, and extends the theories, methods, technologies, and application systems of human intelligence.[1] It may be easily said that, basically, this refers to teaching the technological devices how to learn, perceive, process massive data, and test the data and make decisions. It would not be an exaggeration to state that AI is a machine that behaves like a human. For this reason, AI is also called “cognitive computing”. Ultimately, the indicated deep learning and document automation system proves that technological devices (e.g. computers) are better at their tasks than human beings.

Within the legal industry, there is no doubt that AI is an indispensable agent for all lawyers. Especially, concerning the time-consuming issues in the legal sphere, AI may do such things to surpass that in a way that the ability to operate comparative legal works such as; contracts, documents, as well as all other legal structures. It is obvious that AI may change the way lawyers think and perform their legal work rather than routines.

Reconstructing the Legal Sphere

Artificial intelligence may be used in the judiciary system in a widening framework such as simple assistant to robotic judges who directly make decisions. They can be categorized as follows:

1. Assistant AI:

The most prominent advantage here is that artificial intelligence provides more comprehensive information to the judge/prosecutor/lawyer or related persons in a shorter time.  For example:

• Artificial intelligence can be used effectively to review and analyse many documents. In this direction, documents that a human would need to review for days can be scanned much faster by AI. Even prepare a summary and analysis of the file and it is provided more comprehensive information about its scope. So that the judge will have more comprehensive information regarding the dispute during decision-making process. This comprehensive information also beneficial to lawyers to prepare their own defences especially for the case files.
• AI can conduct much more comprehensive doctrine and Supreme Court decision research aiming at to a dispute. Explaining legal reasoning and legal scientific decisions make easier the proceedings and it also make possible to provide the reliability of the decision. It may be beneficial for both lawyers and judges to prepare their defences and decision-making process either.
• AI may assist in the drafting of the hearing minute during the hearing. It will be very useful for recording witness statements, especially in hearings where many witnesses are heard. This can be beneficial to reduce the length of hearings and minimize the overlapping of court hours.
• AI can assist in drafting decisions in similar cases. This may shorten the time for the preparation of reasoned decisions and their notification to the parties, which is vital in terms of preventing the loss of any right of the parties.

2. Preparation of standardized documents by using AI

Preparation of standardized legal documents such as contracts and power of attorney with the usage of AI may be given as an example. These are already used by many lawyers or law firms. The preparation of documents that are similar to each other in general terms with AI will also both save a time and minimize the possibility of mistake. One of the best things is that AI can quickly scan for new legislation, thus it makes easier to determine whether the contract in question complies with the relevant law or any other regulation. In this manner, documents that can be created as a result of long research can be created in a short time with only few mistakes in high quality.

3. Decision Predicting by AI

AI may be useful in identifying risks and predicting possible outcomes for potential disputes. Here, AI can predict the outcome of a case by analysing the previous decisions made by the judges or the decisions made against similar cases. Within this scope, Lex Machina system developed by Stanford University could be considered as a pioneer in this field. [2] Although these analyses are not binding under the principle of the independence of the judges, this will enable the parties to retain the data to challenge the decision if the AI analysing decisions on similar issues and reaches a conclusion that differs from the court outcome.

4. Direct Decision-Making AI (Robot Judge)

It might be the first thing that comes to mind is using of artificial intelligence in the judiciary as robot judges. Even if it is still very controversial due to the risks that it still harbours within itself for the time being, it may be encountered with robot judges making decisions in the future.

As it is known, DoNotPay, the application developed as the world's first robot lawyer, was to advise the defendant in a court in the US in February. However, CEO and founder Joshua Browder announced that it would be "postponing" those plans, citing "threats from State Bar prosecutors." Even this development can be considered as an indication that society is not yet ready for such developments. Therefore, it is still a question mark whether we will encounter decision-making robots in the near future or not.

Potential Risks & Uncertainties

Despite all, even though saving time is a certain concern, AI usage in law is not ready to fully operate in place of human beings since AI may have an inadequate ability to understand and internalize the logic behind the legal systems. The contemporary national legal systems are generally based on one of four basic law systems which are the civil law, common law, statutory law, religious law or combinations of these.[3] In this direction, it is hard for the AI to predict, use and match the legal systems in a case at hand properly. Besides, there are only limited binding legislative definitions adopted to ensure the legal certainty.[4] In the absence of a legal framework of AI, it will be complicated to provide legal and technological framework.

Apart from that, there may be some peculiar or different ways to referring a subject rather than repeating itself. Furthermore, some documents that include confidential contents and sensitive data are generally protected under law. Within this scope, AI may face some security problems and will not work accurately. Hereby, it may be stated that privacy concerns about the AI will be at the top of the agenda consequently. In addition to privacy concerns, since the AI continues to transform, it is possible to influence the competition market and may leads to inequality between the software companies. For instance, sudden rise of some companies such as Google displacing centuries old ‘great’ companies on the top of the software list is a pointer to competition concerns.[5]

Overmore, concerning AI applications in the legal sphere may result in another risk. As all know, lawyering skill is one of the most remarkable features, and concurrently, the possibility of for AI lawyer replacing lawyer in courtrooms is a distant possibility. In accordance with the vast knowledge, an effective lawyer should have excellent skills in oral advocacy, communication, empathy, negotiating, drafting, and managing. Thus, the poignant thinking way of the lawyer role could not be replaced with the AI technology in courtrooms.

On the other hand, divergently, some argue that AI is designed solely for limited tasks. Should an example be given, AI used for specific areas such as checking the weather conditions or analyzing raw data to calculate the potential parameters etc. However, AI indisputably affects the law systems, and it will be one of the aforementioned specific areas from now on. Conversely, due to convoluted learning process of AI, the ability of develop train itself shall be triggered. Thereupon, it could be argued that therein a grey area for AI to serve perfectly and it needed to be developed and reviewed before being used on a daily basis.

Certainly, the costs of implementation and development of AI algorithms impacted the transition process as well, and most AI systems require human supervision while using the data for implication. To simplify, high-cost studies should be carried out in order to use AI with all the authority for its autonomous feature and mitigate the human intervention to the system in practice.

Conclusion

We may conclude that, as the analysis conducted above demonstrates, legal duties and legal feature are raising complexity in the AI usage in the legal world. Even trusting AI, some risks should be considered when control is passed to it. Nevertheless, may be noted that a new chapter shows up with this autonomous system since AI has various facilities to compare with a human being. In essence, the introduction of the things we have listed under the titles of “Assistant AI” and “Preparation of standardized documents by using AI” can save our time, provide more qualified documents and preliminary preparation. On the other hand, the introduction of “Robot Judges” who decide directly on their own does not seem very possible for today when evaluated together with the risks they will bring.

At this stage, it is paramount that any cooperation between lawyers and technologists shall help for the usage of AI in the legal sphere. Such cooperation may change the requirements of the algorithms and application of the AI.

Ultimately, in order to use AI systems in the judiciary, countries ought to adopt comprehensive regulations for maintaining the AI mechanism in the context of validating the decision-making process and success of the implementation. Even such a step is not easy to forward, AI will be used for law work in the future to create higher standards and thereupon we must follow collaboration for developing the intelligent system both in legal and invention phases.

[1] https://www.researchgate.net/publication/30874496_Artificial_Intelligence_A_Modern_Approach

(Last accessed: 11.10.2022)

[2] For more: https://lexmachina.com/legal-analytics/ (Last accessed: 22.03.2023)

[3] List of national legal systems - Wikipedia (Last accessed: 03.11.2023)

[4] The most relevant are: OECD: recommendations for AI; available at: https://legalinstruments.oecd.org/en/instruments/OECD-LEGAL-0449  (Last accessed: 22.02.2023); European Commission for the Efficiency of Justice (CEPEJ): CEPEJ European Ethical Charter on the use of artificial intelligence (AI) in judicial systems and their environment - European Commission for the Efficiency of Justice (CEPEJ) (coe.int) (Last accessed: 22.02.2023)

[5] Jose, Jomon P, Legal Liability Issues and Regulation of Artificial Intelligence (AI), 2017-18

Pg.36