In the past ten years, the development of artificial intelligence (,,AI’’) has become increasingly prominent in different domains and areas of social, educational and other systems.

Digitalization is not a new term in Serbian dictionary, and artificial intelligence, with its continual improvement, inevitably leads a new ‘’digital revolution’’ whose goal is to improve the standard and quality of life of citizens, increase efficiency, reduce possible costs and loss of time that we can encounter in various situations (for example – in proceedings before administrative bodies).

Because of this reasons, as a result of digitalization, the Republic of Serbia began implementing various models of electronic (online) submissions of documents and applications in order to reduce costs, save time and avoid long waiting queues that can sometimes ‘’stretch’’ even outside the facility where certain institution is located. Also, this segment of digitalization gained more importance in the period of the corona virus epidemic, where, in order to ensure people’s health care, technology was the best solution for the further functioning of state institutions and their interaction with citizens.

Bearing in mind all of the above, as a logical step forward, artificial intelligence, as the latest innovation by digitalization, is trying to find its purpose and importance in today’s modern society.

At the end of 2019, the Republic of Serbia adopted the Strategy for the Development of Artificial Intelligence in the Republic of Serbia for the period 2020-2025 (‘’Official Gazette of the RS’’ no. 95/2019) and in which one of the goals is the improvement of public sector services using artificial intelligence, which is another step towards fulfillment of the conditions for membership in European Union, bearing in mind that states-members of that organization, and beyond, are already widely implementing artificial intelligence in various fields.

However, in this text, we will specifically deal with one aspect of artificial intelligence, specifically its application in law and judicial system, i.e. the term ‘’Predictive justice’’ and what it means.

The term ‘’Predictive justice’’ implies the process of predicting the final outcome or a particular phase of the court procedure by artificial intelligence, based on a previously processed number of court decisions from the same or similar cases.

The first thing that comes to mind when reading such a simplified definition of the term ‘’Predictive justice’’ is that this process of data processing by artificial intelligence relies predominantly on established court practice, which is the main characteristic of precedential law, i.e. countries of the ‘’common law’’ legal system (such as Great Britain and the United States of America, for example).

Therefore, the following question can be asked – how to apply ‘’predictive justice’’ in the Republic of Serbia, as a country that does not belong to the ‘’common law’’ legal system, but to the European-continental system, established during the time of ancient Rome.

The jurisprudence of our country unwaveringly and without hesitation defends the stance that judicial practice in any branch of the law does not represent a source of law. That place is reserved first of all for the constitution, as the supreme and general legal act, then the law, by-laws such as ordinances, regulations, etc. However, it is also the opinion of legal theorists that regardless of the fact that court practice de jure is not a source of law, as it is, for example, in the USA, de facto court practice represents a very relevant,, source of law’’. That means that lower instance courts pay special attention when making decisions, if a similar legal matter has been already resolved by a higher instance court. So, for example, the basic or higher court, in the first instance, although the law does not state this anywhere, will make a decision with special care in a certain legal matter, if the same or identical (only with different parties and in another procedure) has already been decided by the appellate court. Bearing in mind the fact that court practice is not formally a source of law, lower instance courts will not take care of it ex officio, but only if one of the parties submits court practice on its own initiative as one of the pieces of evidence along with the other attachments and evidentiary proposals.

Bearing in mind the above, although the Republic of Serbia is not officially in the system of precedent law, court practice, regardless, gains relevance that is almost equal to the formal source of law.

In comparative law, ‘’predictive justice’’ finds its purpose primarily through two situations:

  • Case management – artificial intelligence is provided with a special algorithm whose purpose is to assess the degree of complexity of the case upon submission of the initial submission that initiates court proceedings, after which it assesses the resources that would be invested in solving it as, well as the extent of the involvement of the court itself, that is competent for a specific legal case;
  • Assistance to the judge in making court decisions – certain federal states of the United States of America, such as California and Wisconsin, use a special software called ‘’COMPAS’’ that evaluates the probability of repeating a criminal offense and thus helps the judge decide on the criminal sanction that would be a substitute for detention.

Besides USA, the European countries that are making the path to the implementation of artificial intelligence in law and judicial system are certainly the Netherlands and Estonia, where in Estonia there is even a special project that has a goal to create robot judge who would be assigned jurisdiction to decide in small value disputes, and if an appeal was filed against the decision of computer judge, it would be decided by a traditional judge.

Artificial intelligence can certainly contribute and help judges in their work, because its own algorithm allows it to process data in a capacity that cannot be achieved by human efforts. Also, this method paves the way for solving many problems faced by our courts, among which is the length of court procedures.

Predictive justice can also, with proper implementation, significantly contribute to the better realization of the principle of legal certainty, by introducing standardization of court decisions with those before the introduction of artificial intelligence and helping to reduce arbitrariness. Judges will be able to pay attention to more complicated cases that require more professional and empirical knowledge, while during that time, ‘’predictive justice’’ would solve, or at least bring to a state before solving, simpler cases that quickly fill the court registries and thus significantly reduce the efficiency of the court.

Of course, all of this would not be possible to implement without a previous special technical procedure through which predictive justice would see a light of day as an idea in our legal system, in which, besides IT experts who would be responsible for the construction and development of algorithms and software, an inevitable role in its progress would also have judges and lawyers who would oversee the proper operation of artificial intelligence, because no matter how much innovation in the field of technology advances, one computer software cannot be programmed to learn everything. The best example would be the application of Article 8 of the Civil Procedure Law regarding the judge’s free assurance in making decisions in dispute and through the perspective of predictive justice, i.e. artificial intelligence and whether the same could be applied at all.

Also, other problems that may arise from the application of predictive justice are of an ethical nature, the issue of neutrality and justice, as well as bearing the responsibility for the decision made, how to teach the program itself such characteristics that are inherent to living beings, not computer programs.

In addition, predictive justice, as such, would naturally imply technology developed by man, and the occurrence of technical errors is quite possible, which would entail certain consequences.

Bearing in mind all these reasons, the European Commission for the Efficiency of Justice (CEPEJ), a body of Council of Europe of which the Republic of Serbia is a member, adopted the European Ethics Charter on the use of artificial intelligence in judicial systems. The Charter prescribes 5 basic principles as a prerequisite for the use of any artificial intelligence in the judiciary of a certain country:

1) Respect for basic rights when designing and applying artificial intelligence;

2) Preventing the development or encouragement of any discrimination against individuals or groups of individuals;

3) Insisting on the quality and security of processes and data;

4) Enabling access and understanding of data processing methods and external audit of those methods;

5) Allowing system users to have enough information and autonomy to be able to understand the process, consider alternatives, including the possibility of having their case decided by a judge.

In the end, there is no doubt that progress in the field of technology can bring a large number of reliefs for the more efficient work of the judicial system of the Republic of Serbia. What is important is that when introducing predictive justice into the judicial system, with great attention, moral and ethical principles, technical process, and protection of basic human rights must be respected.