"I understand your perspective": LLM Persuasion through the Lens of Communicative Action Theory
Esra D\"onmez, Agnieszka Falenska
Why It Matters
What makes this one worth your time
Understanding LLMs' persuasive capabilities can enhance their design for applications requiring nuanced communication, such as negotiation or customer service.
The paper explores LLMs' persuasive abilities through communicative action theory, showing they often outperform humans in conveying intent.
Summary
The paper investigates the persuasive capabilities of Large Language Models (LLMs) using Jürgen Habermas' Theory of Communicative Action, focusing on their ability to express illocutionary intent in online discussions. The study compares LLM-generated counter-arguments with human-written ones, finding that LLMs often convey intent more effectively and are preferred by crowd-sourced workers.
Key contributions
- Application of Habermas' Theory of Communicative Action to evaluate LLMs' persuasive capabilities.
- Comparison of LLM-generated and human-written counter-arguments in terms of illocutionary intent and effectiveness.
Notable insights
- LLMs can convey illocutionary intent more effectively than humans, potentially increasing their anthropomorphism.
- LLMs' sycophantic responses align closely with opinion holders' intents, aiding in opinion change.
Possible limitations
- Not stated in the abstract
Abstract
arXiv:2606.08076v2 Announce Type: replace-cross Abstract: Large Language Models (LLMs) can generate high-quality arguments, yet their ability to engage in nuanced and persuasive communicative actions remains largely unexplored. This work explores the persuasive potential of LLMs through the framework of J\"urgen Habermas' Theory of Communicative Action. It examines whether LLMs express illocutionary intent (i.e., pragmatic functions of language such as conveying knowledge, building trust, or signaling similarity) in ways that are comparable to human communication. We simulate online discussions between opinion holders and LLMs using conversations from the persuasive subreddit ChangeMyView. We then compare the likelihood of illocutionary intents in human-written and LLM-generated counter-arguments, specifically those that successfully changed the original poster's view. We find that all three LLMs effectively convey illocutionary intent -- often more so than humans -- potentially increasing their anthropomorphism. Further, LLMs craft sycophantic responses that closely align with the opinion holder's intent, a strategy strongly associated with opinion change. Finally, crowd-sourced workers find LLM-generated counter-arguments more agreeable and consistently prefer them over human-written ones. These findings suggest that LLMs' persuasive power extends beyond merely generating high-quality arguments. On the contrary, training LLMs with human preferences effectively tunes them to mirror human communication patterns, particularly nuanced communicative actions, potentially increasing individuals' susceptibility to their influence.