The Judicial Review of the Automated Administrative Act

More and more, administrative authorities are using algorithms to improve the efficiency of public decision making. Although algorithms are supposed to guarantee the neutrality, speed and adequacy of the decision, they nevertheless give rise to questions, which are now submitted to the judge. But is jurisdictional review, designed to control the human decision, adapted to a decision taken on the basis of an algorithm?
This question, largely unexplored, has been raised on the occasion of a more general study on the use of digital technologies by administrations and (administrative) judges, a study proposed by the French Society of Comparative Legislation and the University of Leiden*. On the basis of a questionnary sent at the end of 2018 to their correspondents throughout the world, the national responses to the theme of judicial review of decision- making processes based on algorithms were rather elliptical, due to the lack of litigation in the States surveyed. However, the disputes that have been brought to our attention through this survey are very instructive.
Litigation has crystallized on questions relating to the legal foundations of control. In this respect, certain French and Dutch cases offer particularly interesting avenues for reflection, revealing major trends on two issues: The question of communication of the algorithm is the one that comes up most often (I), as well as the question of its supervision (II).