Detecting Conversational Mental Manipulation with Intent-Aware Prompting

TL;DR

This paper introduces Intent-Aware Prompting (IAP), a novel method leveraging large language models to detect covert mental manipulation in conversations by understanding underlying participant intents, showing improved accuracy over existing strategies.

Contribution

The paper presents a new IAP approach that enhances detection of mental manipulation by capturing participant intents, outperforming existing prompting methods in effectiveness.

Findings

IAP significantly reduces false negatives in manipulation detection.

Experimental results show superior performance of IAP on the MentalManip dataset.

The approach helps identify more manipulation instances with minimal false positives.

Abstract

Mental manipulation severely undermines mental wellness by covertly and negatively distorting decision-making. While there is an increasing interest in mental health care within the natural language processing community, progress in tackling manipulation remains limited due to the complexity of detecting subtle, covert tactics in conversations. In this paper, we propose Intent-Aware Prompting (IAP), a novel approach for detecting mental manipulations using large language models (LLMs), providing a deeper understanding of manipulative tactics by capturing the underlying intents of participants. Experimental results on the MentalManip dataset demonstrate superior effectiveness of IAP against other advanced prompting strategies. Notably, our approach substantially reduces false negatives, helping detect more instances of mental manipulation with minimal misjudgment of positive cases. The code of this paper is available at https://github.com/Anton-Jiayuan-MA/Manip-IAP.