Automated Social Science: Language Models as Scientist and Subjects NBER

An NBER working paper presenting an approach for automatically generating and testing social scientific hypotheses in silico, using large language models with structural causal models. The approach is demonstrated in a negotiation, a bail hearing, a job interview, and an auction.

Website Directory import
0 shouts Added Sep 23, 2026 nber.org
www.nber.org/papers/w32381
Automated Social Science: Language Models as Scientist and Subjects NBER website preview

Links

Tags