Judicious use of LLMs could speed up progress in the social sciences

NEWS AND VIEWS

08 July 2026 Judicious use of LLMs could speed up progress in the social sciences Large language models could advance experiments in the social sciences, but fundamental questions remain about how they should be used to support research. By James N. Druckman 0 James N. Druckman James N. Druckman is in the Department of Political Science, University of Rochester, Rochester, New York 14627, USA. View author publications PubMed Google Scholar

Social scientists spend considerable time exploring how people reason and what affects their opinions. However, designing and implementing good experiments to probe these questions is time-consuming and costly. Writing in Nature, Ashokkumar et al.1 report that large language models (LLMs) can be used to forecast the outcomes of social-science experiments. The authors’ findings suggest that these models could be used as low-cost tools to assist experimental design and thereby accelerate advances in social science — but the findings also raise questions about how best to use the models effectively and responsibly.

doi: https://doi.org/10.1038/d41586-026-01875-0

References Ashokkumar, A., Hewitt, L., Ghezae, I. & Willer, R. Nature https://doi.org/10.1038/s41586-026-10742-x (2026). Druckman, J. N. Experimental Thinking (Cambridge Univ. Press, 2022). Chu, J. Y. et al. Proc. Natl Acad. Sci. USA 121, e2307008121 (2024). Zellner, M., Abbas, A. E., Budescu, D. V. & Galstyan, A. R. Soc. Open Sci. 8, 201187 (2021). Bryan, C. J., Tipton, E. & Yeager, D. S. Nature Hum. Behav. 5, 980–989 (2021). Bisbee, J., Clinton, J. D., Dorff, C., Kenkel, B. & Larson, J. M. Polit. Anal. 32, 401–416 (2024). Download references

Competing Interests The author declares no competing interests.

Subjects