Abstract This article examines the relationship between artificial intelligence and legal decision-making through the lens of computability. It argues that many challenges associated with AI in law do not arise solely from technical limitations, but from structural features of legal reasoning itself. Drawing on computability theory, the article shows that legal judgment cannot be fully reduced to algorithmic procedures, as it relies on interpretation, proportionality, and discretionary justification. These elements constitute a non-computable core of law that resists exhaustive formalization. The analysis then explores how AI-mediated decision-making reshapes institutional structures through predictive rationality, algorithmic opacity, and hybrid forms of agency. In response, the article develops a due process framework grounded in contestability, meaningful human oversight, and accountable hybridity. It concludes by proposing a limits-aware approach to AI governance, in which computational systems support but do not replace the interpretative practices that sustain legal legitimacy.
Salvatore Stanizzi (Thu,) studied this question.