When Do LLMs Need Human Experts? Evidence for Social Science from Jurisprudential Classification
A new study finds that even frontier large language models like GPT-5.2 cannot replace expert human annotation for complex social science tasks, as they underperform fine-tuned BERT when classifying l…