Alignment as Jurisprudence: Bridging AI Ethics and Legal Theory
A new academic paper titled "Alignment as Jurisprudence," published in the Yale Journal of Law and Technology and archived on arXiv, explores the structural parallels between legal jurisprudence and artificial intelligence alignment. The author, Nicholas Caputo, argues that both fields aim to predict and shape the decisions of powerful actors—judges in law and AI models in technology—through the specification and interpretation of language. The essay integrates Ronald Dworkin’s principle-oriented interpretivism and Cass Sunstein’s positivist analogical reasoning with modern AI techniques like Constitutional AI and case-based reasoning. It suggests that legal debates can offer insights for improving AI alignment, while lessons from AI development can enhance understanding of legal processes. The core argument posits that both systems should empower human capabilities and goal achievement. As AI capacities grow and traditional legal constraints on judges evolve, the interdisciplinary conversation becomes increasingly vital for refining both legal theory and AI safety frameworks, pointing toward improved versions of both disciplines.
Wire timeline
Alignment as Jurisprudence: Bridging AI Ethics and Legal Theory
A new academic paper titled "Alignment as Jurisprudence," published in the Yale Journal of Law and Technology and archived on arXiv, explores the structural parallels between legal jurisprudence and artificial intelligence alignment. The author, Nicholas Caputo, argues that both fields aim to predict and shape the decisions of powerful actors—judges in law and AI models in technology—through the specification and interpretation of language. The essay integrates Ronald Dworkin’s principle-oriented interpretivism and Cass Sunstein’s positivist analogical reasoning with modern AI techniques like Constitutional AI and case-based reasoning. It suggests that legal debates can offer insights for improving AI alignment, while lessons from AI development can enhance understanding of legal processes. The core argument posits that both systems should empower human capabilities and goal achievement. As AI capacities grow and traditional legal constraints on judges evolve, the interdisciplinary conversation becomes increasingly vital for refining both legal theory and AI safety frameworks, pointing toward improved versions of both disciplines.
cs.AI updates on arXiv.org