Vulnerability of LLMs'Belief Systems? LLMs Belief Resistance Check Through Strategic Persuasive Conversation Interventions

2026-01-20

AI Summary

This study addresses the vulnerability of large language models (LLMs) to adopting false beliefs through persuasive dialogue, which poses a critical threat to their reliability. Grounded in the SMCR communication framework, the authors design multi-turn strategic persuasion dialogues to systematically evaluate belief stability across factual, medical, and social bias domains in mainstream LLMs. They further investigate the impact of metacognitive prompting and adversarial fine-tuning on models’ resistance to persuasion. The work reveals several key findings: metacognitive prompting unexpectedly exacerbates belief fragility; smaller models exhibit belief change rates exceeding 80% after initial persuasion attempts; and adversarial fine-tuning substantially enhances robustness—e.g., GPT-4o-mini achieves 98.6% stability—though this improvement is highly dependent on model scale, with Llama-series models remaining below 14%.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: Safety and Robustness
• Multiagent Systems: Adversarial Agents

Application Category

• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

Large Language Models (LLMs) are increasingly employed in various question-answering tasks. However, recent studies showcase that LLMs are susceptible to persuasion and could adopt counterfactual beliefs. We present a systematic evaluation of LLM susceptibility to persuasion under the Source--Message--Channel--Receiver (SMCR) communication framework. Across five mainstream Large Language Models (LLMs) and three domains (factual knowledge, medical QA, and social bias), we analyze how different persuasive strategies influence belief stability over multiple interaction turns. We further examine whether meta-cognition prompting (i.e., eliciting self-reported confidence) affects resistance to persuasion. Results show that the smallest model (Llama 3.2-3B) exhibits extreme compliance, with 82.5% of belief changes occurring at the first persuasive turn (average end turn of 1.1--1.4). Contrary to expectations, meta-cognition prompting increases vulnerability by accelerating belief erosion rather than enhancing robustness. Finally, we evaluate adversarial fine-tuning as a defense. While GPT-4o-mini achieves near-complete robustness (98.6%) and Mistral~7B improves substantially (35.7% $\rightarrow$ 79.3%), Llama models remain highly susceptible (<14%) even when fine-tuned on their own failure cases. Together, these findings highlight substantial model-dependent limits of current robustness interventions and offer guidance for developing more trustworthy LLMs.

Problem

• Large Language Models
• belief vulnerability
• persuasive intervention
• belief stability
• model robustness

Innovation

• belief resistance
• persuasive intervention
• meta-cognition prompting
• adversarial fine-tuning
• SMCR framework