The Illusion of Intervention: Your LLM-Simulated Experiment is an Observational Study
Researchers have identified a critical flaw in using large language models (LLMs) to simulate human behavior for experimental studies. Because LLMs are trained on observational data, interventions can inadvertently alter the simulated users' underlying attributes, leading to "user drift." This drift can distort the estimated effects of interventions, making the experimental results unreliable. The study proposes methods to diagnose this confounding using negative control outcomes and mitigate it by adjusting LLM personas with relevant confounders. AI
IMPACT Highlights a potential pitfall in using LLMs for experimental research, impacting the reliability of findings in behavioral science and AI studies.