BLTG: Sebastian Reyes Molina - Augmented Justice: A Case of Legitimate Authority скачать в хорошем качестве

BLTG: Sebastian Reyes Molina - Augmented Justice: A Case of Legitimate Authority

Belgrade Legal Theory Group (June 16, 2025) Prof. Sebastian Reyes Molina "Augmented Justice: A Case of Legitimate Authority" Belgrade Legal Theory Group hosted a lecture by Prof. Sebastián Reyes Molina (Maastricht University), on the topic of artificial intelligence and legal adjudication. The event was part of the summer semester series for 2025. Prof. Reyes opened with a discussion on the weaknesses of human judges in the administration of justice. Judges, he argued, are vulnerable to a range of limitations—bias, fatigue, political pressure, and limited computational capacity. These shortcomings have led some to believe that AI might offer a better alternative. To explore this possibility, he presented a thought experiment based on a fictional government initiative called “Augmented Justice”, which uses advanced AI systems to create avatars modelled on human judges – Hercules fully replaces human judges, while Socrates provides advisory opinions that can be accepted or rejected by the court. The program is said to achieve just decisions 99% of the time, with justice defined as the accurate determination of facts and the correct interpretation of legal sources. Prof. Reyes then addressed a series of plausible objections (legitimacy, morality, bias, and accountability) that challenge the acceptability of AI even if it produces just decisions – as distinct from putative objections, which question whether AI can truly replicate legal reasoning. Prof. Reyes argued that if legitimacy is based on the ability to make substantiated decisions based on legal knowledge, then AI can qualify. On the moral side, critics claim that AI lacks the capacity for moral knowledge and judgment, whereas a counter-objection points out that moral knowledge is a problematic notion in itself. He also addressed the so-called “black box” and bias problems, noting that human judges often struggle with similar issues, including implicit bias and lack of transparency. Regarding accountability, Prof. Reyes acknowledged that AI cannot be punished or removed like human judges. However, he noted that in the Socrates model, final responsibility remains with the human court, and in the Hercules model, it might be possible to design legal frameworks to address the accountability gap. The lecture ended with what Prof. Reyes called the Adjudicator’s Puzzle: if AI reliably makes just decisions, why do we remain uneasy about giving it authority? Is justice simply the outcome of good reasoning, or is there something more required from the adjudicator? The discussion that followed touched on the accuracy of the assumptions, the process of adjudication and the role of AI, as well as the philosophical foundations of adjudication. At the end of the meeting, Sava Vojnović thanked everyone for an engaging discussion, as well as Prof. Reyes for his excellent lecture.