July 23, 2025

LLMs in the Lab: Can AI Predict What Real Participants Do?

View Full Paper

Ask AI

Bookmark

LLMs can replicate the effect direction and significance found in real human data from randomized controlled trials.
The study analyzed multiple LLMs, including ChatGPT and Gemini, across various experimental domains.
Findings indicate that LLM-simulated datasets provide insights for refining study designs and enhancing research robustness.
This approach offers a collaborative tool, complementing traditional empirical studies without replacing them.

Discussion

No takes yet. Share an insight, caveat, or question.

Experimental analysis confirms LLMs can mirror real RCT outcomes, suggesting potential in study design refinement.

Lukumon et al. (2025) studied this question.

View Full Paper

Ask AI

Bookmark