Home - IVR 2024
AI and the Rule of Law
Vincent Chiao (University of Toronto, Faculty of Law) vincent.chiao@utoronto.ca
François Tanguay-Renaud (Osgoode Hall Law School) ftanguay-renaud@osgoode.yorku.ca
Aziz Huq (University of Chicago School of Law) huq@uchicago.edu
Boris Babic (University of Toronto; Hong Kong University) boris.babic@utoronto.ca
Sandra Wachter (Oxford University) sandra.wachter@oii.ox.ac.uk
Boris Babic, Assistant Professor, University of Toronto; Associate Professor, Data Science, Law and Philosophy, Hong Kong University
Vincent Chiao, Associate Professor, University of Toronto
Aziz Huq, Frank and Bernice J. Greenberg Professor of Law, University of Chicago
François Tanguay-Renaud, Associate Professor, Osgoode Hall Law School
Sandra Wachter, Professor of Technology and Regulation, University of Oxford

Widespread use of AI, fueled by advances in foundation models such as ChatGPT and Dall-E, have the potential to both challenge and reshape our conceptions of the rule of law and its value. On the one hand, they challenge the basic tenets of modern legal systems, based on clearly stated and transparent rules and public accountability in court. On the other hand, they also provide an occasion for reshaping our understanding of the rule of law’s value in a technologically advanced society.
AI capabilities are increasing at a rapid clip and are poised to be widely deployed across a diverse range of commercial and governmental contexts. However, AI systems can operate in ways that are opaque and difficult or impossible to interpret, making it difficult to justify decisions based on their outputs in traditional legal fora. Moreover, the dynamic character of AI systems can unsettle expectations of predictability and consistency, fostering suspicions that decisions based on AI models are arbitrary, opaque, and unfair. Consequently, AI decision-making poses a challenge to the rule of law, whether understood (as Dicey understood it) primarily as a matter the legal accountability of public officials, or as a stable, clear, and accessible rules-based legal order (as Fuller argued we should understand it.)
At the same time, however, our reasons for valuing the rule of law are rooted in the rule of law’s ability to mitigate human tendencies toward partiality, in-group favoritism, inconsistency, and self-aggrandizement. A rules-based legal order that insists upon the impartial and uniform application of general rules by public authorities is a form of social technology created to manage these tendencies. Seen in this light, the development of powerful, context-sensitive AI tools—which may not be subject to the same types of epistemic and practical limitations—may be as much a complement as it is a threat to the rule of law.
This panel on the theme of AI and the rule of law will focus on these challenges and opportunities. Panelists will focus on challenges for the regulation and governance of foundation models; assessing the justifiability of a decision independent of its intelligibility; conceptualizing arbitrariness of AI outputs in high-stakes legal settings; as well as how the rule of law may both transform, and be transformed by, the use of AI in law.