Modelli predittivi a supporto della decisione giudiziaria. Alcuni spunti di riflessione

This essay focuses on some legal and epistemological implications linked to the use of artifi cial intelligence in judicial decisions. The authors try to highlight the link amongst scientific data, statistics and artificial intelligence with the aim to investigate how this information can be used by a judge in his/her decisional process. By analyzing the main studies concerning the use of technology in trials and reconstructing the school-case known as the Loomis case, the article tries to understand if the artificial intelligence can ensure the certainty of law with a correct and fair decision or if it can never replace totally the legal reasoning led by a human agent (the judge). The essay arises from a research led within the national project “LAILA – Legal Analytics for Italian Law” (P.R.I.N. 2017).