Humans and LLMs Split Theory-Building Strengths
A blinded comparison of 25 LLMs and 73 scientists finds stronger individual AI performance but more prediction-efficient, aggregable human theories.
Sep 29, 20264 min2609.32562
Computing's societal impact, ethics, policy, and educational implications.
A blinded comparison of 25 LLMs and 73 scientists finds stronger individual AI performance but more prediction-efficient, aggregable human theories.
NOHARM uses specialist-rated consult options across 1,100 cases, finding severe-harm potential up to 24.6% and gains from AI-assisted physicians.