GPT-5 outperforms federal judges 100% to 52% in legal reasoning experiment
Summary
The SSRN preprint reports that GPT-5 achieved a 100% accuracy rate in a legal reasoning experiment, allegedly outperforming federal judges (52%). The article emphasizes potential implications for AI-assisted legal work while noting the need for methodological details before drawing firm conclusions. Readers should scrutinize the experimental design, prompts, datasets, and evaluation criteria.