Sign up to save tools and stay up to date with the latest in AI
bg
bg
1

AI Judges Follow The Law, Human Judges Follow Their Hearts, Study Reveals

Mar 20, 2025 - forbes.com
A study by University of Chicago Law School researchers Eric A. Posner and Shivam Saran compared AI and human judicial decision-making, revealing significant differences. Using OpenAI's GPT-4o to replicate an experiment with 31 U.S. federal judges, the study found that AI adhered to legal precedent over 90% of the time, unaffected by the defendant's likability. In contrast, human judges were influenced by sympathetic portrayals of defendants 65% of the time, often deviating from precedent. Law students showed intermediate behavior, following precedent 85% of the time. This highlights the divide between legal formalism, which AI embodies by strictly following rules, and legal realism, which human judges demonstrate by considering extralegal factors.

Efforts to make AI incorporate emotional factors like human judges were unsuccessful, suggesting a deep-rooted difference in reasoning. The study underscores a philosophical debate in legal philosophy: whether justice should be blind and rule-based or consider human elements and context. While AI offers consistency and predictability, human judges bring compassion and understanding, reflecting the complexity of justice. The research suggests that the difference between AI and human judicial reasoning is more philosophical than technological, leaving open the question of which approach better serves justice.

Key takeaways:

  • The study highlights a stark contrast between AI and human judicial decision-making, with AI adhering strictly to legal precedent while human judges are influenced by sympathy.
  • Human judges deviated from legal precedent in 65% of cases when faced with sympathetic defendants, whereas AI followed precedent in over 90% of cases.
  • The research underscores the ongoing debate between legal formalism and legal realism, with AI embodying the former and human judges demonstrating the latter.
  • Attempts to make AI incorporate emotional factors like human judges were unsuccessful, suggesting a deep-rooted difference in judicial reasoning between AI and humans.
View Full Article

Comments (0)

Be the first to comment!