Recommendations for the use of AI in the Judiciary (Part I)

For this reason, one of the discussions in research on AI and law is the proposal for a guideline that adequately informs the jurisdictions about the use of this technology in their case.

The European Commission for the Efficiency of Justice (CEPEJ), created in 2002, has the objective of improving the efficiency and the functioning of the Judiciary in the Member States, through the definition of instruments, metrics of evaluation and elaboration of documents. CEPEJ is responsible, among other tasks, for analyzing the results of the judicial systems, identifying the difficulties encountered by them, promoting assistance and establishing concrete means to improve the evaluation of the results and the functioning of these systems.

In this context, in December 2018, it adopted the document “European Ethical Charter on use of artificial intelligence in judicial systems”, which announces a set of principles capable of guiding public policy makers, legislators, legal professionals, public and private actors responsible for design and development of Artificial Intelligence (AI) systems involving the processing of judgments and judicial data.

This written instrument is structured according to premises that encourage the use of AI in the Judiciary, conformity with the rules of protection of human rights and protection of personal data, respect for the fundamental rights of people and the requirements of transparency, impartiality, equity, evaluation by independent experts and certified regularity.

This is explained, according to CEPEJ, by the multiplicity of uses of AI systems in the judicial space. In civil, commercial and administrative law, the processing of judicial decisions can, for example, contribute with predictability in the application of the law, uniformity and consistency in the decisions of the Courts. However, in criminal matters, the risk of discrimination linked to the use of an unrepresentative data set in addition to the processing of sensitive data, must be mitigated, in order to guarantee a fair trial.

For this reason, one of the discussions in research on AI and law is the proposal for a guideline that adequately informs the jurisdictions about the use of this technology in their case, which implies transmitting the characteristics of the decision-making process by AI, risks and consequences, so that the subject can consciously choose to authorize this use or not.

This written instrument is divided into the presentation of the principles, perception of the state of use of AI in the Judiciary of the Member States, possible applications derived from AI in the European judicial systems and glossary.

The principles are structured around five axes: (1) respect for fundamental rights; (2) non-discrimination; (3) quality and safety; (4) transparency, impartiality and equity; (5) preponderance of user control.

The possible applications derived from AI are divided into: (1) uses to be encouraged (jurisprudence, access to justice, creation of new strategic tools); (2) applications that require methodological precautions (elaboration of scales in civil litigations, support for appropriate methods of conflict resolution, “online dispute resolution”, use of algorithms in criminal investigation); (3) uses that require scientific studies (analysis of the decisions of each judge, predictive analysis of judicial decisions); and (4) uses to be considered with the most extreme reservations (algorithms in criminal matters to create profiles of individuals, provide the synthesis of all decisions already taken as a basis to be used in future decisions).

By: Wilson Sales Belchior