Upotreba veštačke inteligencije za procenu rizika od recidivizma u krivičnom postupku i poštovanje prava na pravično suđenje iz člana 6. stav 1. Evropske konvencije o ljudskim pravima

In contemporary criminal justice systems, complex artificial intelligence (AI) models are increasingly utilized to assess the risk of recidivism, thereby rendering crime prevention instruments algorithmic in nature. The paper examines the concept of recidivism as a form of crime prevention and the instruments employed for assessing recidivism risk within criminal proceedings, with particular emphasis on tools that leverage AI and machine learning to assist judges in decision-making processes. Such AI models are already in use in the United States. Taking into account the issues that have emerged in practice during the application of these models in the U.S., as well as the general challenges recognized in the use of AI, this paper analyses these problems from the perspective of upholding the guarantees of the right to a fair trial under Article 6, paragraph 1 of the European Convention on Human Rights. The objective is to determine whether such models, in their current form, can be implemented in the member states of the Council of Europe in accordance with the accepted standards of fairness in criminal proceedings.