KMI: A Dataset of Korean Motivational Interviewing Dialogues for Psychotherapy

2025-02-08

AI Summary

High-quality training data for Motivational Interviewing (MI) in non-English contexts is scarce, and existing AI-based mental health chatbots suffer from limitations in privacy preservation, clinical fidelity, and cultural adaptation. Method: We introduce KMI—the first theoretically grounded, professional-grade synthetic Korean MI dialogue dataset comprising 1,000 dialogues—generated via an LLM-driven controllable dialogue synthesis pipeline that deeply integrates core MI principles, clinician knowledge injection, MI-specific behavioral prediction models, and advanced prompt engineering. We further propose a novel MI-theory-informed dialogue evaluation metric and a multi-expert blind review framework for rigorous validation. Contribution/Results: Experiments demonstrate that KMI significantly improves downstream models’ performance in empathic response generation and MI strategy accuracy. KMI establishes a reproducible, evaluable data foundation and methodological paradigm for multilingual AI systems in mental health.

Technology Category

• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval
• Natural Language Processing: Code Generation / Program Synthesis from Natural Language
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

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

Abstract

The increasing demand for mental health services has led to the rise of AI-driven mental health chatbots, though challenges related to privacy, data collection, and expertise persist. Motivational Interviewing (MI) is gaining attention as a theoretical basis for boosting expertise in the development of these chatbots. However, existing datasets are showing limitations for training chatbots, leading to a substantial demand for publicly available resources in the field of MI and psychotherapy. These challenges are even more pronounced in non-English languages, where they receive less attention. In this paper, we propose a novel framework that simulates MI sessions enriched with the expertise of professional therapists. We train an MI forecaster model that mimics the behavioral choices of professional therapists and employ Large Language Models (LLMs) to generate utterances through prompt engineering. Then, we present KMI, the first synthetic dataset theoretically grounded in MI, containing 1,000 high-quality Korean Motivational Interviewing dialogues. Through an extensive expert evaluation of the generated dataset and the dialogue model trained on it, we demonstrate the quality, expertise, and practicality of KMI. We also introduce novel metrics derived from MI theory in order to evaluate dialogues from the perspective of MI.

Problem

• AI-driven mental health chatbots
• Motivational Interviewing dataset limitation
• Korean language MI dataset development

Innovation

• Simulates MI with therapist expertise
• Uses LLMs for utterance generation
• Introduces novel MI evaluation metrics