Stanford Study Shows AI Answers Preferred Over Law Professors' Responses

A Stanford Law School-led study evaluated nearly 3,000 anonymized answers to contract‑law questions from 16 law professors and large‑language‑model AI systems. Professors favored the AI‑generated responses in about 75% of the head‑to‑head comparisons, rating them higher for accuracy, reasoning, clarity and educational value. The AI answers were also flagged as pedagogically harmful far less often (≈3.5%) than professor‑written answers (≈12%). The research, involving scholars from Stanford, Yale, NYU, Chicago and other institutions, suggests AI could serve as an effective supplemental tutor in law courses, though the authors stopped short of recommending replacement of faculty.

The study focused on contracts law because it requires nuanced interpretation rather than simple factual recall. Participants created representative student questions, supplied their own answers, and compared them with AI‑generated counterparts calibrated to similar length and structure. Findings highlight the potential for on‑demand, individualized assistance, especially in large classes or institutions with limited faculty resources.