Overstating Attitudes, Ignoring Networks: LLM Biases in Simulating Misinformation Susceptibility

TL;DR

This study evaluates how well large language models can simulate human susceptibility to misinformation, revealing they overstate certain associations and ignore network factors, thus serving better as diagnostic tools than substitutes.

Contribution

It demonstrates the limitations of LLMs in accurately replicating human misinformation susceptibility patterns and highlights their biases and systematic distortions.

Findings

LLMs capture broad distributional tendencies but overstate belief-sharing associations.

Models show higher explained variance and emphasize attitudinal features over network data.

Distortions stem from biases in how misinformation concepts are represented in training data.

Abstract

Large language models (LLMs) are increasingly used as proxies for human judgment in computational social science, yet their ability to reproduce patterns of susceptibility to misinformation remains unclear. We test whether LLM-simulated survey respondents, prompted with participant profiles drawn from social survey data measuring network, demographic, attitudinal and behavioral features, can reproduce human patterns of misinformation belief and sharing. Using three online surveys as baselines, we evaluate whether LLM outputs match observed response distributions and recover feature-outcome associations present in the original survey data. LLM-generated responses capture broad distributional tendencies and show modest correlation with human responses, but consistently overstate the association between belief and sharing. Linear models fit to simulated responses exhibit substantially higher explained variance and place disproportionate weight on attitudinal and behavioral features, while largely ignoring personal network characteristics, relative to models fit to human responses. Analyses of model-generated reasoning and LLM training data suggest that these distortions reflect systematic biases in how misinformation-related concepts are represented. Our findings suggest that LLM-based survey simulations are better suited for diagnosing systematic divergences from human judgment than for substituting it.