On the Adaptive Psychological Persuasion of Large Language Models

2025-06-07

AI Summary

This study presents the first systematic investigation of large language models’ (LLMs’) dual capabilities—autonomous persuasion and resistance to persuasion—within psychological rhetorical contexts. Addressing the limitations of existing methods, namely poor strategy generalizability and insufficient situational adaptability, we propose an adversarial dialogue-based evaluation framework coupled with a context-aware adaptive strategy selection mechanism. Our approach integrates 11 empirically grounded psychological persuasion techniques (e.g., fluency effect, repetition effect) and employs direct preference optimization (DPO) for dynamic, fine-grained strategy tuning. We validate the framework on three leading open-source LLMs, demonstrating statistically significant improvements in persuasion success rates while preserving baseline general-purpose capabilities. All code is publicly released to support reproducibility and further research.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: (Large) Language Models
• 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

Previous work has showcased the intriguing capabilities of Large Language Models (LLMs) in instruction-following and rhetorical fluency. However, systematic exploration of their dual capabilities to autonomously persuade and resist persuasion, particularly in contexts involving psychological rhetoric, remains unexplored. In this paper, we first evaluate four commonly adopted LLMs by tasking them to alternately act as persuaders and listeners in adversarial dialogues. Empirical results show that persuader LLMs predominantly employ repetitive strategies, leading to low success rates. Then we introduce eleven comprehensive psychological persuasion strategies, finding that explicitly instructing LLMs to adopt specific strategies such as Fluency Effect and Repetition Effect significantly improves persuasion success rates. However, no ``one-size-fits-all'' strategy proves universally effective, with performance heavily dependent on contextual counterfactuals. Motivated by these observations, we propose an adaptive framework based on direct preference optimization that trains LLMs to autonomously select optimal strategies by leveraging persuasion results from strategy-specific responses as preference pairs. Experiments on three open-source LLMs confirm that the proposed adaptive psychological persuasion method effectively enables persuader LLMs to select optimal strategies, significantly enhancing their success rates while maintaining general capabilities. Our code is available at https://github.com/KalinaEine/PsychologicalPersuasion.

Problem

• Evaluating LLMs' persuasion and resistance in adversarial dialogues
• Improving persuasion success with psychological strategies
• Developing adaptive framework for optimal strategy selection

Innovation

• Evaluates LLMs as persuaders and listeners
• Introduces eleven psychological persuasion strategies
• Proposes adaptive framework for optimal strategy selection