ChatThero: An LLM-Supported Chatbot for Behavior Change and Therapeutic Support in Addiction Recovery

2025-08-28

AI Summary

Over 36 million people globally suffer from substance use disorders (SUDs), yet treatment accessibility and efficacy remain severely limited by stigma, low patient motivation, and insufficient personalized support. To address this, we propose the first clinically aligned multi-agent dialogue framework for addiction recovery, integrating cognitive behavioral therapy (CBT) and motivational interviewing (MI). We construct a high-fidelity, resistance-graded synthetic benchmark and adopt a two-stage training paradigm—supervised fine-tuning followed by direct preference optimization—incorporating dynamic patient modeling, context-aware dialogue generation, and adaptive persuasive strategies. Experiments demonstrate that our system increases average patient treatment motivation by 41.5% and self-efficacy by 0.49 points, reduces resolution cycles for challenging cases by 26% compared to GPT-4o, and significantly outperforms existing models in empathy, responsiveness, and behavioral authenticity. Clinical validity is confirmed through both automated and expert clinician evaluations.

Technology Category

• Multiagent Systems: Adversarial Agents
• Cognitive Modeling & Cognitive Systems: Adaptive Behavior
• Humans and AI: AI for Accessibility

Application Category

• User Modeling, Personalization and Recommendation: Metrics for user behavior and evaluating success
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Search and Retrieval-Augmented AI: Agentic search

Abstract

Substance use disorders (SUDs) affect over 36 million people worldwide, yet few receive effective care due to stigma, motivational barriers, and limited personalized support. Although large language models (LLMs) show promise for mental-health assistance, most systems lack tight integration with clinically validated strategies, reducing effectiveness in addiction recovery. We present ChatThero, a multi-agent conversational framework that couples dynamic patient modeling with context-sensitive therapeutic dialogue and adaptive persuasive strategies grounded in cognitive behavioral therapy (CBT) and motivational interviewing (MI). We build a high-fidelity synthetic benchmark spanning Easy, Medium, and Hard resistance levels, and train ChatThero with a two-stage pipeline comprising supervised fine-tuning (SFT) followed by direct preference optimization (DPO). In evaluation, ChatThero yields a 41.5% average gain in patient motivation, a 0.49% increase in treatment confidence, and resolves hard cases with 26% fewer turns than GPT-4o, and both automated and human clinical assessments rate it higher in empathy, responsiveness, and behavioral realism. The framework supports rigorous, privacy-preserving study of therapeutic conversation and provides a robust, replicable basis for research and clinical translation.

Problem

• Addressing substance use disorders with personalized therapeutic chatbot support
• Integrating clinically validated strategies into LLM-based addiction recovery systems
• Overcoming motivational barriers and treatment resistance in addiction care

Innovation

• Multi-agent framework with dynamic patient modeling
• Two-stage training with SFT and DPO optimization
• Integrates CBT and motivational interviewing strategies