The Anthropological Sovereignty of Law: Artificial Intelligence, Interpretation, and the Human Reserve of Judgment in the Algorithmic Age

Abstract

The expansion of artificial intelligence systems constitutes one of the foremost legal challenges of the twenty-first century. This article advances an original theoretical category: the principle of the anthropological sovereignty of law. According to this principle, final legal decisions producing legal effects, or comparably significant effects on fundamental rights, must remain attributable to a competent and accountable human decision-maker acting within the framework of law. Artificial intelligence may assist that decision-maker, but it cannot become an autonomous or parallel source of decisional authority. This requirement rests on the distinctively human capacity to understand the existential significance of legal conflict and to assume responsibility for its resolution. Two corollaries follow: the semantic irreducibility of law, according to which legal meaning cannot be reduced to computational output, and the human reserve of judgment, which gives the principle its institutional and procedural dimension. The article explains and defends the use of the term sovereignty in contrast to related concepts such as dignity, human centrality, and the primacy of the person. It then tests the thesis against some of the strongest arguments in favour of extensive reliance on artificial intelligence in legal decision-making, including Kahneman, Sibony and Sunstein’s theory of noise and Casey and Niblett’s proposal of microdirectives. The analysis subsequently applies this framework to Regulation (EU) 2024/1689 (the AI Act), interpreted as the most systematic legislative articulation to date, within European law, of the requirement of human control over algorithmic decision-making. The principle is then examined in relation to the principal legal professions — advocate, judge, civil-law notary and legal adviser — in order to identify what remains irreducibly human in the exercise of each function. The concluding section addresses the objection that the proposed category merely renames legal personalism. It argues, instead, that the principle of the anthropological sovereignty of law translates the axiological core of personalism into an operational legal category capable of responding to an interlocutor that traditional personalism never had to confront: artificial intelligence.