Towards an LLM-Based Multilingual CBT Chatbot for Anxiety Management: A Comparative Study and Preliminary Development

This study introduces a multilingual chatbot based on Large Language Models (LLMs) and Cognitive Behavioral Therapy (CBT) designed to assist with anxiety management in clinical settings. The proposed system integrates an adaptive, context-aware chatbot can provide of delivering therapeutic dialogue in multiple languages. The LLM-based chatbot supports cognitive restructuring by engaging users in a structured and empathetic manner. Patients interact with the multilingual chatbot during supervised sessions via a secure hospital web application. Preliminary evaluations highlight the chatbot’s potential for scalable, ethical, and comprehensive digital mental health interventions, demonstrating promising results in response latency, emotional understanding, and therapeutic coherence.