Abstract

The possibility of using AI systems for the purpose of limiting the excesses of self-referentiality sometimes produced by a stable body of judges is analysed. In addition, the question of whether the problem of correct motivation of a judicial decision in the case of confrontation between indigenous cultures and Western legal systems can be solved with machine learning algorithms is addressed. Regarding the first question, an artificial system can control the excesses of the judiciary’s self-referentiality under some conditions. For the second question, the experiment carried out with machine learning on applications for asylum raises some further questions. It is shown that the judgment is based more emotionally than rationally, due to the difficulties of translating meanings between cultures far apart in evolutionary history. The failure of networks to learn supports this hypothesis and confirm the importance of the judge’s purely emotional evaluation in order to reach a decision