Contenu du sommaire : Les défis de la justice numérique

The notion of artificial intelligence is described from four perspectives: its aims, its relationship to computation, its investigation methods, and the knowledge it produces. These factors are described and then a general hypothesis elaborated on the relations between artificial intelligence and information technology: the aim of artificial intelligence is to confirm the alphanumeric vision of the world deployed by information technology.

How might artificial intelligence (AI) be useful to justice and what does it take? While reducing the complexity of procedures is a real concern, information technology needs are not the same for everyone. While some types of AI have already proven their worth in practice, will robots ever judge in place of men as prophesied twenty years ago? At present, artificial intelligence can help litigants, litigants and judges to organise information and enrich procedures, but there is still a long way to go before it meets the fair trial standards set by the ECHR.

Knowledge of court production presupposes an analysis of the representative corpus of judicial decisions. The promise of automated analysis raises issues of feasibility and reliability, and questions as to the normative role of the knowledge arising therefrom. From an operational perspective, in terms of reducing the uncertainty inherent in the operation of the justice system, we must also account for the choice - even the automatic choice - of what criteria will be used as legitimate determinants of the decisions that are handed down. The fact that the reference frameworks are established must not obviate an examination of their construction.

The aim of the Interactive Visualization of Court of cassation Jurisprudence (VICO) project is to create an interactive and visual representation of the jurisprudence on the website of the Court of cassation (https://www.data.gouv.fr/fr/datasets/cass/).In this article, we have chosen, in a first part, to take an historical approach to visualization as applied to mediaeval law in order to show its importance in the construction of knowledge, and in a second part to make a presentation of the interactive jurisprudence visualization project.

The application of technological innovation to the field of justice does hold promise, but is also a subject of debate and even a source of concern. It is both a tool for streamlining institutional action, and a means for guaranteeing access to information by the public. But are the massive investments being made by most European countries in this domain proof of the quality of the decisions that are ultimately handed down? To analyse this, it is necessary to examine the role and scope of artificial intelligence in the judicial sphere.

The notion of "predictive" justice that has appeared in recent years in the judicial landscape generally refers to the process of applying the latest automatic big-data analysis techniques to case law databases, with the stated goal of predicting the outcome of trials to the greatest possible extent. The irrational fears and unrealistic expectations the notion evokes in some must be kept at a distance, in order to allow for a calm assessment of its real potential and of the possible risks posed. These conditions will make possible the gradual development of the outlines of a reasonable and virtuous use of this technology, which is now still in its infancy but will very likely experience a rapid development.

Artificial Intelligence is currently a major topic of discussion, but it is often poorly defined. Here we will seek to go back to the foundations of how it works: mathematical and statistical formalism. That kind of formalism, which performs well in quantifiable, closed environments, is far less suited to open environments such as the justice system. The exploitation of jurisprudence with these means does not allow legal briefs to be faithfully taken into consideration, as processed rulings do not necessarily contain all the causative elements of a decision. This criticism must also not obscure another reality, which cuts across the whole of our society: a project is taking shape now to bring about government by data, which is likely to pose a challenge to the very foundations of our democracies, including the rule of law.

Christophe Jamin here presents his translation of a text, previously unpublished in France, by Jerome Frank (1899-1957), a legal expert who was one of the founders of the American school of legal realism. After practicing for twenty years as a business lawyer in Chicago and New York, he became one of the major figures of the New Deal. In this text, and in his controversial book "Law & the Modern Mind," published in 1930, the author attacks the "basic legal myth" that judges never create a right but simply deduce it from clear, definite, and invariable premises. For Frank, the real life of the law is made of uncertainties, gaps, and contradictions, and psychological, and - more broadly - extra-legal considerations play a major role.

The name of Pouvanaa a Oopa, a Polynesian separatist political leader, former member of parliament and senator, sentenced in 1959 to eight years'imprisonment for attempted arson of the city of Papeete, was finally cleared more than forty years after his death. The case was brought before the Court of review and re-examination of criminal convictions by the Minister for Justice in June 2014, and the deceased leader's name was cleared by a judgment promulgated on 25 october 2018 ; this had a great impact in French Polynesia, where the ruling was highly anticipated, but went nearly unnoticed in mainland France. The aim of this article is to return to this case, which strongly marked the history of relations between France and Polynesia, in order to expose the process that led to this act of legal review, in many ways an exceptional one, and to situate the case in a broader context of the legal review of criminal convictions in an attempt to measure its legal, political and symbolic scope.

The justice system, which creates legal rights with long-term effects for citizens, is making efforts to increase its effectiveness by digitizing the flow of information it generates while guaranteeing that the reliability and integrity of that information will be preserved, in a manner capable of indisputably producing the legal effects prevailing upon their production. The judicial archives, logically, have been assembled under the concurrent effect of these two requirements. The aim of this article is thus to provide a synthetic overview of those procedures situated within the sphere of activity of the Ministry for Justice that have already led to the production of digital data, and publicize the rules governing its access, dissemination and processing that allow researchers to exploit those resources.

In matters concerning educational assistance, juvenile judges may feel that they are participating in the public monitoring of poverty. While it is true that powerful mechanisms of domination do exist in this domain, judges can nevertheless take account of the poverty of the individual by acting in accordance with specific principles in their interactions with the parties to the cases brought before them. Between the dual pitfalls of naivety and a merely abstract denunciation of domination, methodical discomfort assumes the judge's awareness of the need to take ownership of the rule of law in a strategic manner.