Judicial Artificial Intelligence

In 1938, the automatic calculating machine was used in the House of Representatives of the State of Ohio. This was one of the first times technology was used in the legal environment. However, it has not been extensively studied until the American jurist Loevinger published his study in 1949 on 'Jurimetrics', in which the concept of Jurimetrics was created as a reference to the application of automation to contribute to the rationalisation of law, raising the possibility of foreseeing legal decisions on empirically based calculations. Later in 1961, the American Association of American Law Schools Jurimetrics Committee on Scientific Investigation of Legal Problems put out a report saying that in the field of jurimetrics, they were most interested in things like using machines to analyse the quantitative parts of legal decisions. Therefore, it's clear that the idea of using technology to help judges make decisions was already being thought about at that time. It is worth noting that in 2018, professors at the University of Southern California and Stanford Law School published research revealing that artificial intelligence is more accurate than lawyers when it comes to contract review. In India, the Supreme Court of India has set up an Artificial Intelligence Committee in 2022 to look into how AI can be used in the legal domain.

Taking into account of this background, currently, what is understood under the term artificial intelligence refers to a large number of advanced techniques for the mathematical processing of data based on the idea of automation, such as deep learning or artificial neural networks, can be highlighted for the specific case. The latter are techniques based on the biological structure of the human brain, seeking to adapt and reproduce some of its capacities, such as the mental process that judicial operators, such as judges, go through when making a decision. Thus, the application of artificial intelligence in the field of judicial decisions is often referred to as 'Judicial Artificial Intelligence.'

In order to analyse the legal problems that may arise as a result of the application of artificial intelligence in judicial decisions, it is necessary to understand at a general level how these tools work. The first thing is that they work under the concept of predictive analytics, where behavioural patterns are provided. It is a tool based on the analysis of singular big data, which is used to try to predict future decisions based on the systematic study of decisions already produced.

These mechanisms are based on predictive algorithms implemented to carry out actions or recommendations, starting from an existing set of data, based on patterns and probabilities. The entire context and programming is carried out by human beings, who, by virtue of the information they adapt and the algorithms they generate to relate the variables, determine the predictive behaviour of the machine in its entirety. In this way, from a group of data, patterns and relationships between variables are determined to establish a prediction regarding a certain future behaviour.

The information used for the programming of the expert system requires the joint collaboration of the programmer and the legal analysts, since the quality and quantity of the information will depend on what both consider appropriate, taking into account the complexity of the kind of problems to be developed, requiring a considerable amount of knowledge about the area to be dealt with. It should also be noted that expert systems need to be constantly updated with new information, so that they can provide predictions in line with the state of affairs at any given time.

In relation to judicial decisions, the information with which expert systems must be programmed must be fully in tune with the legal system of a country, which involves the regulations at different levels, i.e. laws, decrees, regulations, circulars, among others, as well as rulings of the main courts and tribunals, which allow a complete analysis to be made at the time of the problem under study. These predictive analytics mechanisms operate under expert legal systems, which help judges make decisions by proposing possible solutions to given legal issues, applying the expert knowledge embedded in their system to a given case.

An expert system is a set of logical artificial intelligence procedures that allow computers to support humans in analysing situations and making decisions. They are then databases that have not only knowledge, but also a set of conditional rules that link knowledge in an operational way. At a general level, therefore, an expert system differs from a database in which the data is statically stored, whereas in expert systems the data is dynamic, thus not only pointing out the information, but also linking it to the cases to be dealt with; in this way, the information is processed as a human being who is an expert in that field would do.

With respect to the specific case, legal expert systems are programmes that are suited to act as intelligent legal operators, aiming to find solutions to legal problems presented to them by means of the information contained in their databases. Lastly, it should be taken into account that expert legal systems need, in order to provide their solution, to have contact with the user, which in most cases is done through a system of questions where the system receives the necessary information to carry out the appropriate processing that allows it to provide an answer. Thus, the expert system will be able to advise the user as long as the latter interacts with the programme by providing it with guidelines on which it has to make a decision.

With what has been said so far, it can be concluded firstly that, as has been insisted, the aim of applying artificial intelligence to law is not for expert systems to make the decision, but to help the judicial branch to have all the regulatory and jurisprudential elements applicable in a decision, speeding up the investigation process, as well as basing them on purely objective questions.

Therefore, by basing judicial decisions on objective matters, without excessive influence of aspects of human nature such as states of mind, beliefs, etc., it is possible to guarantee greater impartiality and accuracy in the decision itself, which increases confidence in judges and gives legal certainty to the public in general, thereby increasing the favourability of the judiciary.

Finally, the responsibility for the decisions that are taken must rest with the judge, who is the one who, with his own reasoning, makes the decision that he considers appropriate in the specific case, without this being attributed to the creators or designers or to the tool itself. However, designer must take responsibility when creating the system to ensure that it is in line with the principles of law.