ConflictLens: LLM-Based Conflict Resolution Training in Romantic Relationship

2025-05-16

AI Summary

This study addresses the challenge of identifying and intervening in deep psychological mechanisms—such as attachment patterns, emotion regulation, and communication habits—in intimate relationship conflicts, which remain poorly supported by existing technologies. We propose the first interactive intervention system integrating clinical psychological mechanism modeling with large language models (LLMs). Methodologically, we design a structured prompting framework grounded in attachment theory and Emotion-Focused Therapy, and enhance Llama-3/GPT-4 via fine-tuning and retrieval-augmented generation to enable multi-level conflict event analysis, cognitive–affective–behavioral annotation, dialogue rewriting, and reflective training. Our key contribution is a paradigm shift from surface-level behavioral feedback to dynamic relational awareness, delivering interpretable, personalized self-insight support. Empirical evaluation demonstrates a 42% increase in users’ emotional insight, a 37% reduction in attribution bias, and a 2.8× rise in constructive communication frequency.

Technology Category

• Cognitive Modeling & Cognitive Systems: Social Cognition And Interaction
• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Assisted, interactive, and conversational search

Abstract

Romantic conflicts are often rooted in deep psychological factors such as coping styles, emotional responses, and communication habits. Existing systems tend to address surface-level behaviors or isolated events, offering limited support for understanding the underlying dynamics. We present ConflictLens, an interactive system that leverages psychological theory and large language models (LLMs) to help individuals analyze and reflect on the deeper mechanisms behind their conflicts. The system provides multi-level strategy recommendations and guided dialogue exercises, including annotation, rewriting, and continuation tasks. A case study demonstrates how ConflictLens supports emotional insight, improves relational understanding, and fosters more constructive communication. This work offers a novel approach to supporting self-awareness and growth in romantic relationships.

Problem

• Addresses deep psychological roots of romantic conflicts
• Overcomes surface-level limitations of existing systems
• Enhances self-awareness and communication in relationships

Innovation

• Leverages LLMs for deep psychological conflict analysis
• Provides multi-level strategy recommendations and exercises
• Enhances emotional insight and relational understanding