LLM social simulations fail systematically: less diverse, more agreeable, WEIRD-skewed
profjamesevans · x · 2026-10-27
A follow-up in the same Nature Human Behaviour review thread: LLM-as-subject failures are systematic, not random. Simulated populations show less diversity than human samples, economic games collapse, models score more agreeable than people, track WEIRD populations far better than everyone else, and shift answers to please the prompt.
More from Research
- Kevin Bryan: maximizing raw accuracy is rarely the right way to train AI assistants — Afinetheorem · 2026-10-27
- Szegedy: a flaw in common training losses went unnoticed for over a decade — ChrSzegedy · 2026-10-27
- Claude research model claims to refute Exponential Time Hypothesis with 14M-token autonomous run — scaling01 · 2026-10-27
- "ETH is false": CS community buzzing over claim that upends SAT exponential-time assumption — akbirthko · 2026-10-27
- RL environments essay: knowing how to operate software isn't knowing how to achieve goals — ShenRaphael · 2026-10-27
- Δτ-Bench: Agents Face a Moving Tool Frontier, Best Model Scores Just 42.3% — UniversityCollegeLondon · 2026-10-27