Law, Justice, Reason-Giving, and Artificial Intelligence
How do judges, ordinary people, and AI make decisions? Eyal Zamir draws on behavioral science to explore when judges do not strictly apply legal norms, how emotions shape our intimate choices, and whether artificial intelligence reasons like humans.
In a recently published article (with Ori Katz), Eyal Zamir demonstrated that requiring legal decision-makers to provide reasons for their decisions increases adherence to formal legal rules when those rules conflict with intuitions of justice. He now extends this inquiry to AI. He finds that large language models, such as ChatGPT and Claude, are substantially more formalistic than humans and are hardly affected by reason-giving requirements. He also investigates how reason-giving affects intermediate outcomes between fully accepting or entirely dismissing a claim, using surveys, vignette experiments, and observational analyses.
Turning from institutional decision-making to intimate ones, a second project offers a behavioral analysis of family law. It examines how cognitive biases, emotions, and altruistic motivations shape intimate decisions and interact with legal rules. The goal is to use these behavioral insights to improve substantive family law, as well as decisions about marriage and divorce.
Bringing these strands together at the level of theory, Doron Teichman and Eyal Zamir are substantially revising the first edition of their co-authored book, Behavioral Law and Economics (OUP, 2018), updating it in light of recent significant developments in behavioral science, artificial intelligence, and public policy.