Simulating Human Behavior with AI Agents

https://hai.stanford.edu/policy/simulating-human-behavior-with-ai-agents

1,052 participants; agents replicated GSS responses 85% as accurately as participants replicated their own answers two weeks later; comparable on Big Five, five economic games (dictator, two trust games, public goods, prisoner's dilemma), and five experiments; less biased than earlier simulation methods.

LLM models of individual human choice