Nicholas Caputo studies how law can govern artificial intelligence, and how AI can in turn renew and improve our legal and administrative institutions.

Caputo’s work begins from the problem of how to keep powerful, expert, and inscrutable decision-makers under legitimate democratic authority in their exercise of discretion. This is a problem that the law has long wrestled with and that frontier AI poses acutely. He draws on jurisprudence, constitutional and administrative law, and advancing technical methods in AI alignment, evaluation, and control, using each of these fields as design materials for the others.

His recent research runs along three connected lines. First, his legal alignment work argues that AI alignment and law share a common structure, such that legal theory and methods can inform how AI systems are trained to reason and act (Legal Alignment for Safe and Ethical AI, Transactions on Machine Learning Research (TMLR); Alignment as Jurisprudence, Yale Journal of Law and Technology). Second, he studies the technological history of public and private law, drawing lessons to help adapt our existing institutions to emerging technologies (“Quiet Enjoyment”, Stanford Technology Law Review; Administrative Law’s Fourth Settlement: AI and the Capability-Accountability Trap, working paper). And third, he examines novel private governance instruments and institutions such as AI constitutions and catastrophic risk frameworks, exploring questions of legitimacy and accountability (Can Claude Consent to its Own Constitution? AI Constitutionalism and the Paradox of Constituent Power, working paper).

Caputo has presented his work in technical venues, including workshops at NeurIPS and CHAI, and has advised policymakers around the world on AI governance. He has experience with strategic litigation and policy advocacy in the United States and in the European Union. 

Caputo is also Law & AI Lead at the Oxford Martin AI Governance Initiative (AIGI). He has a J.D. from Harvard Law School and a B.A. from Carleton College. Before law school, he taught debate in China for two years, largely during the COVID-19 pandemic. He speaks French and Mandarin Chinese.