The surprising case for AI judges

Bridget McCormack, former Chief Justice of the Michigan Supreme Court and current CEO of the American Arbitration Association (AAA), discusses the development of the AI Arbitrator, a platform designed to automate dispute resolution, starting with document-only construction disputes. McCormack argues that the current human-led legal system is often unpredictable, inefficient, and lacks public trust, citing high rates of error in lower courts and the inaccessibility of justice for those who cannot afford lawyers. She suggests that AI can improve fairness by ensuring parties feel heard and understood, as the system can transparently show its work, unlike many human judges. The AI Arbitrator uses a series of agents to parse claims and evidence, drafts an award, and always keeps a human arbitrator in the loop for final issuance. While acknowledging risks like AI hallucinations, McCormack emphasizes that their current system is narrowly focused, governed, and audited, performing well against human baselines in testing. She believes that while criminal and government-related cases should remain in public courts, offering AI-assisted arbitration for private disputes provides a necessary, efficient, and potentially fairer option for parties seeking resolution.

(Source：The Verge)