Perils and opportunities in using large language models in psychological research

Abstract

The emergence of large language models (LLMs) has sparked considerable interest in their potential application in psychological research, mainly as a model of the human psyche or as a general text-analysis tool. However, the trend of using LLMs without sufficient attention to their limitations and risks, which we rhetorically refer to as "GPTology", can be detrimental given the easy access to models such as ChatGPT. Beyond existing general guidelines, we investigate the current limitations, ethical implications, and potential of LLMs specifically for psychological research, and show their concrete impact in various empirical studies. Our results highlight the importance of recognizing global psychological diversity, cautioning against treating LLMs (especially in zero-shot settings) as universal solutions for text analysis, and developing transparent, open methods to address LLMs' opaque nature for reliable, reproducible, and robust inference from AI-generated data. Acknowledging LLMs' utility for task automation, such as text annotation, or to expand our understanding of human psychology, we argue for diversifying human samples and expanding psychology's methodological toolbox to promote an inclusive, generalizable science, countering homogenization, and over-reliance on LLMs.

Keywords: large language models; natural language processing; psychological diversity; psychological text analysis; psychology.

Fig. 1.

ChatGPT vs. human moral judgments. Note: a) Distributions of moral judgments of humans (light blue) and GPT (light red) in six moral domains. Dashed lines represent averages. b) Inter-correlations between moral values in humans ($N=\mathrm{3,902}$) and ChatGPT queries ($N=1,000$). c) Network of partial correlations between moral values based on a diverse sample of humans from 19 nations (30) and 1,000 queries of GPT. Blue edges represent positive partial correlations and red edges represent negative partial correlations.

Fig. 2.

Comparing ChatGPT against humans grouped by political opinion for responses on the Big Five Inventory. Note: Figure shows the response distribution of humans and ChatGPT across the five-factor personality constructs and for different human demographics. Figure shows that ChatGPT gives significantly higher responses on Agreeableness, Conscientiousness and significantly lower responses on Openness and Neuroticism. Importantly, ChatGPT shows significantly less variance compared with all demographic groups on all personality dimensions.

Fig. 3.

Comparing ChatGPT against humans across various demographic variables for the Right-Wing-Authoritarianism scale. Note: Figure shows the response distribution of humans and ChatGPT on the RWA scale for different human demographics. ChatGPT shows significantly lower average RWA than male, white, and young participants but not explicitly liberal participants. Importantly, ChatGPT shows significantly less variance compared with all demographic groups.