AI Psychological Support Systems: Research on the Contribution Mechanism of Dialogue Strategies to Immediate Interaction Quality

This project investigates how conversational strategies used by AI-based psychological support systems influence the quality of human–AI interactions. Drawing on Hill’s Helping Skills Model, the study analyzes real conversation transcripts between users and AI tools to identify relationship-, insight-, and action-oriented strategies. Using large language models for large-scale dialogue coding, the research examines how these strategies relate to indicators of interaction quality, such as emotional change, depth of experiencing, and perceived session impact. The findings aim to inform the development of evidence-based AI psychological support systems.