Large Language Models Persuade Without Planning Theory of Mind

2026-02-18

AI Summary

This study addresses the limitation of current Theory of Mind (ToM) evaluations, which predominantly rely on static question-answering and fail to capture large language models’ (LLMs’) capacity to reason about others’ mental states in dynamic interactions. The authors propose the first interactive persuasion task, wherein agents must strategically disclose information to guide a target participant toward selecting one of three policy proposals, distinguishing between conditions where mental states are explicitly provided versus those requiring active inquiry. Through behavioral experiments, role-playing scenarios, and belief measurements, the study compares human and LLM performance under varying information visibility. Results show that while LLMs excel when mental states are explicit, they perform below chance in hidden conditions requiring active inference. Nevertheless, in real human interactions, LLMs demonstrate superior overall persuasive efficacy, revealing their ability to influence others’ beliefs through rhetorical strategies without explicit ToM reasoning.

Technology Category

• Cognitive Modeling & Cognitive Systems: Social Cognition And Interaction
• Natural Language Processing: Discourse, Pragmatics & Argument Mining
• Knowledge Representation and Reasoning: Reasoning with Beliefs

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

A growing body of work attempts to evaluate the theory of mind (ToM) abilities of humans and large language models (LLMs) using static, non-interactive question-and-answer benchmarks. However, theoretical work in the field suggests that first-personal interaction is a crucial part of ToM and that such predictive, spectatorial tasks may fail to evaluate it. We address this gap with a novel ToM task that requires an agent to persuade a target to choose one of three policy proposals by strategically revealing information. Success depends on a persuader's sensitivity to a given target's knowledge states (what the target knows about the policies) and motivational states (how much the target values different outcomes). We varied whether these states were Revealed to persuaders or Hidden, in which case persuaders had to inquire about or infer them. In Experiment 1, participants persuaded a bot programmed to make only rational inferences. LLMs excelled in the Revealed condition but performed below chance in the Hidden condition, suggesting difficulty with the multi-step planning required to elicit and use mental state information. Humans performed moderately well in both conditions, indicating an ability to engage such planning. In Experiment 2, where a human target role-played the bot, and in Experiment 3, where we measured whether human targets' real beliefs changed, LLMs outperformed human persuaders across all conditions. These results suggest that effective persuasion can occur without explicit ToM reasoning (e.g., through rhetorical strategies) and that LLMs excel at this form of persuasion. Overall, our results caution against attributing human-like ToM to LLMs while highlighting LLMs' potential to influence people's beliefs and behavior.

Problem

• Theory of Mind
• Large Language Models
• Persuasion
• Interactive Task
• Mental State Inference

Innovation

• interactive theory of mind
• strategic persuasion
• mental state inference
• large language models
• multi-step planning