Artificial intelligence and the reasoning of judges
- By Etienne Vergès
- and Géraldine Vial
Pages 205 to 221
Cite this article
- VERGÈS, Etienne
- and VIAL, Géraldine,
- Vergès, Etienne.
- et al.
- Vergès, E.
- and Vial, G.
https://doi.org/10.3917/cdlj.2502.0205
Cite this article
- Vergès, E.
- and Vial, G.
- Vergès, Etienne.
- et al.
- VERGÈS, Etienne
- and VIAL, Géraldine,
https://doi.org/10.3917/cdlj.2502.0205
The reasoning processes of judges are at the heart of research on artificial intelligence tools. While initial studies focused on modeling the decision-making processes of judges and the influence of predictive tools on their rulings, the development of generative AI has marked a significant scientific and technical turning point. AIs now simulate human cognitive functions and produce legal reasoning. In this study, we seek to test their ability to understand a legal situation and make an appropriate decision, up to and including the drafting of a judgment. The study thus reveals certain legal skills of AIs, although the tools still exhibit numerous reliability and security flaws.