 Droit et intelligence artificielleAurore Hyde (),
Agnes Delaborde (),
Guillaume Bernard,
Bénédicte Girard (),
Sophie Harnay (),
Christian Licoppe () and
Liza Veyre ()
Working Papers from HAL Abstract: The research focuses on the regulation of predictive justice tools. Technically, these tools are computer applications that use more or less advanced machine learning and Natural Language Processing technologies to analyse a large volume of previous decisions. These technologies are used to reproduce, by machine, the work of analysing judicial decisions that a human being could have done, to a lesser extent, in order to carry out a gigantic task of sorting and analysing their text. By making the state of litigation on a given issue visible on a large scale, these tools go well beyond a simple documentary research function and are intended to influence, ultimately, the decision of the legal professional who uses them. In addition, by giving prominence to decisions by trial judges that were previously relatively unknown, predictive justice tools influence the ordering of sources of law and create a risk of ‘factualisation' of the law. Lastly, they have an impact on the guarantees arising from the right to a fair trial, such as effective access to an independent and impartial judge and equality of arms. Under these conditions, the way in which they are designed and used cannot be left to the laws of the market alone. The aim of our research is therefore to sketch out a global approach – combining economics, law, metrology and sociology – to the question of how to guarantee confidence in the public justice service in the light of predictive justice tools. After demonstrating the need to regulate the market for predictive justice tools, the work carried out in the law-economics area was designed to examine the methods of this regulation. Both the methods of ex ante regulation of predictive justice tools, by analysing the modern approach to compliance in particular, and the methods of ex post regulation, by looking at the responsibility of players in this area, were studied. The aim of the research in the law and metrology area was to complement the theoretical approach to compliance by studying the feasibility of evaluating predictive justice tools with a view to their possible certification. This work, carried out in partnership with experts from the Laboratoire national de métrologie et d'essais (LNE), enabled an initial set of specifications to be drawn up for conducting a risk analysis and an assessment plan based on a certifiable standard. The aim of the law and sociology section was to consider, from both a practical and theoretical perspective, how predictive justice tools can change the way the law is made and justice delivered. The work carried out highlighted the need for professional training in this area and the need to prioritise decisions on the merits using a method that has yet to be developed. At the end of the report, it emerged that academic research and public authorities in the broadest sense have a crucial role to play in ensuring that these technologies genuinely improve access to justice while preserving the fundamental principles of law. Keywords: Intelligence Artficielle; Justice prévisionnelle; Numérique; Open data; Régulation (search for similar items in EconPapers) Published in 20.21, IERDJ - Institut des Études et de la Recherche sur le Droit et la Justice. 2025, pp.336 Downloads: (external link) Related works: Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text Persistent link: https://EconPapers.repec.org/RePEc:hal:wpaper:hal-05169360 Access Statistics for this paper More papers in Working Papers from HAL |
Is your work missing from RePEc? Here is how to contribute.
Questions or problems? Check the EconPapers FAQ or send mail to .
EconPapers is hosted by the
School of Business at Örebro University.