The Artificial Intelligence and the dispute for different ways in its predictive use in the criminal process

The article explores the dispute between two global models about the acceptance of artificial intelligence to carry out predictive research around judges and courts, with special focus on criminal procedures. While in the United States research is free and has been developing, in France there has been recent criminalization of the behavior of those who use judicial decisions to do so. Given the delicate cooperation that must exist between experts in criminal procedure and knowledge engineers in the selection and construction of the algorithms that will teach the machine for the preparation of the predictive research, which one would be appropriate? The disparity of interpretative premises in the criminal process put in risk some possible syllogistic constructions that may arise from entimemas? Does the use of artificial intelligence in predictive analysis of criminal procedural decisions violate or ensure constitutional guarantees? Taking these issues into account, the article aims to evaluate whether the US liberal model would be the best option, without forgetting some alerts already identified by the European Comission for the effectiveness of Justice in its recent "European Ethical Charter on the Use of Artificial Intelligence in Judicial Systems and its Environment", notably regarding the risks of prevalence of prejudices that may misrepresent predictive research and its results.