Digital Dictionary

P

Predictive Justice

Bruno Deffains

Université Paris 2 Panthéon-Assas, France

Creating an algorithm capable of correctly solving legal problems is a major objective of legaltech. The work on how the legal system could be automated, with the aim of improving the organization of justice, is old. As early as 1949, Lee Loevinger proposed the application of quantitative methods to the field of justice, an approach he called “jurimetrics”. According to Loevinger, while jurisprudence is based on an approach in which legal reasoning is exclusively a matter of interpreting norms, jurimetrics, on the other hand, uses scientific methods to identify arguments relevant to the law. Loevinger believes that many legal issues could be resolved by applying predictive analysis methods based on systematic processing of court data.

In the context of this “jurimetric” approach, the objective is not so much to know if it is possible to predict the judicial decision as to know how to make such “predictions” by means of a machine, on a quantitative and not intuitive basis. It is worth noting that while legal science has often placed strong emphasis on empirical knowledge in its discourse, there are paradoxically few simple and easily accessible tools for providing reliable statistical data on the application of rules to specific cases. The use of statistics is often confined to the construction of performance indicators, such as those relating to the length of trials or the average amount of compensation ...