You Can't Fool Us: Understanding the Resilience of LLM-driven Agent Communities to Misinformation

2026-05-17

AI Summary

This study investigates how large language model–driven synthetic communities resist and recover from exposure to credible misinformation. By jointly modeling “active open-minded thinking” (AOT) and “political ideology” (PI), the authors construct multi-agent communities exhibiting cognitive and identity heterogeneity, enabling systematic tracking of trust, skepticism, correction, and retraction behaviors during misinformation diffusion. The work evaluates the efficacy of interventions including fact-checking, persuasion, accuracy prompts, and source warnings. It reveals, for the first time, the micro-level mechanisms through which psychological traits and ideology jointly shape community resilience: high AOT significantly enhances both initial resistance and post-exposure recovery; ideologically moderate communities achieve more complete recovery, whereas polarized communities retain greater residual support for misinformation; fact-checking and persuasion prove most effective after the misinformation peak, while accuracy prompts primarily bolster early-stage caution.

Abstract

Misinformation resilience is a dynamic community process: communities differ not only in whether they initially trust false claims, but also in how they recover through interaction, questioning, correction, and support withdrawal. We study this process with an LLM-based agent simulation that constructs synthetic communities along two theoretically motivated dimensions: Actively Open-minded Thinking (AOT), which captures evidence-seeking and willingness to revise beliefs, and Political Ideology (PI), which captures identity-based interpretation of contested claims. These two traits allow us to examine how evidence-oriented reasoning and ideological alignment jointly shape community responses to credible misinformation shocks. Across systematically varied AOT-PI communities, we find that higher AOT improves both resistance to misinformation uptake and recovery after trust peaks. PI shapes the recovery pathway: ideologically moderate communities recover more reliably, while polarized communities retain more residual support. Stance-level analysis shows that resilience depends on whether agents move from questioning a claim to denying or correcting it and withdrawing prior support. Intervention experiments further show that persuasion and fact checking better support post-peak correction, whereas accuracy prompts mainly induce early caution and source warnings have weaker effects. Together, this work provides a mechanism-level account of community misinformation resilience, showing how psychological composition and intervention design shape whether communities move from misinformation exposure toward correction or persistent support.

Problem

• misinformation resilience
• LLM-driven agent communities
• Actively Open-minded Thinking
• Political Ideology
• community recovery

Innovation

• LLM-based agent simulation
• misinformation resilience
• Actively Open-minded Thinking
• Political Ideology
• intervention design