Abstract:Mental health has increasingly become a global issue that reveals the limitations of traditional conversational psychotherapy, constrained by location, time, expense, and privacy concerns. In response to these challenges, we introduce SoulSpeak, a Large Language Model (LLM)-enabled chatbot designed to democratize access to psychotherapy. SoulSpeak improves upon the capabilities of standard LLM-enabled chatbots by incorporating a novel dual-memory component that combines short-term and long-term context via Retrieval Augmented Generation (RAG) to offer personalized responses while ensuring the preservation of user privacy and intimacy through a dedicated privacy module. In addition, it leverages a counseling chat dataset of therapist-client interactions and various prompting techniques to align the generated responses with psychotherapeutic methods. We introduce two fine-tuned BERT models to evaluate the system against existing LLMs and human therapists: the Conversational Psychotherapy Preference Model (CPPM) to simulate human preference among responses and another to assess response relevance to user input. CPPM is useful for training and evaluating psychotherapy-focused language models independent from SoulSpeak, helping with the constrained resources available for psychotherapy. Furthermore, the effectiveness of the dual-memory component and the robustness of the privacy module are also examined. Our findings highlight the potential and challenge of enhancing mental health care by offering an alternative that combines the expertise of traditional therapy with the advantages of LLMs, providing a promising way to address the accessibility and personalization gap in current mental health services.

What problem does this paper attempt to address?

The problems that this paper attempts to solve mainly focus on how to improve the limitations of traditional psychotherapy conversations through large - language - model (LLM) technology. Specifically, the paper proposes solutions in the following aspects: 1. **Geographical, Temporal and Cost Constraints**: Traditional psychotherapy is restricted by geographical location, time arrangement and high cost, resulting in many people in need of psychological support being unable to obtain effective treatment. The paper proposes an LLM - based psychotherapy chatbot named SOULSPEAK, aiming to provide low - cost and easily accessible psychological counseling services via the Internet. 2. **Privacy Protection**: When using a chatbot for psychotherapy, user privacy protection is an important issue. SOULSPEAK enhances users' trust in the system by introducing a special privacy module to ensure the security and anonymity of user information. 3. **Personalized Responses**: In order to improve the therapeutic effect of the chatbot, SOULSPEAK integrates a dual - memory component, combines short - term and long - term contexts, and provides more personalized responses through Retrieval - Augmented Generation (RAG) technology. This enables the system to better understand users' needs and backgrounds, and thus provide more targeted support. 4. **Professionalism and Accuracy**: To ensure that the generated responses meet the professional standards of psychotherapy, SOULSPEAK utilizes a tutoring chat data set containing data of interactions between therapists and clients, and makes the generated responses consistent with psychotherapy methods through multiple prompting techniques. 5. **Evaluation and Optimization**: The paper also introduces a method named "Conversational Psychotherapy Preference Model" (CPPM) to simulate human preferences for different responses, thereby helping to train and evaluate language models related to psychotherapy. In addition, the paper also explores the effectiveness of the long - term memory module and the robustness of the privacy module. In summary, the main goal of this paper is to develop a chatbot system capable of providing high - quality, low - cost psychotherapy services by integrating privacy protection, dual - memory mechanisms and domain - specific expertise using large - language - model technology, in order to address the current problems of insufficient accessibility and personalization in mental health services.

Advancing Conversational Psychotherapy: Integrating Privacy, Dual-Memory, and Domain Expertise with Large Language Models

Enhancing AI-Driven Psychological Consultation: Layered Prompts with Large Language Models

LLM-based Conversational AI Therapist for Daily Functioning Screening and Psychotherapeutic Intervention via Everyday Smart Devices

ChatCounselor: A Large Language Models for Mental Health Support

SMILE: Single-turn to Multi-turn Inclusive Language Expansion via ChatGPT for Mental Health Support

Domain-Specific Improvement on Psychotherapy Chatbot Using Assistant

Enhancing Mental Health Support through Human-AI Collaboration: Toward Secure and Empathetic AI-enabled chatbots

Supporting the Demand on Mental Health Services with AI-Based Conversational Large Language Models (LLMs)

Can AI Relate: Testing Large Language Model Response for Mental Health Support

SoulChat: Improving LLMs' Empathy, Listening, and Comfort Abilities through Fine-tuning with Multi-turn Empathy Conversations

PsyChat: A Client-Centric Dialogue System for Mental Health Support

Enhancing user experience and trust in advanced LLM-based conversational agents

VCounselor: A Psychological Intervention Chat Agent Based on a Knowledge-Enhanced Large Language Model

LLM-empowered Chatbots for Psychiatrist and Patient Simulation: Application and Evaluation

A Novel Nuanced Conversation Evaluation Framework for Large Language Models in Mental Health

Psy-LLM: Scaling up Global Mental Health Psychological Services with AI-based Large Language Models

The Role of Humanization and Robustness of Large Language Models in Conversational Artificial Intelligence for Individuals With Depression: A Critical Analysis

Interactive Agents: Simulating Counselor-Client Psychological Counseling via Role-Playing LLM-to-LLM Interactions

Understanding the Benefits and Challenges of Using Large Language Model-based Conversational Agents for Mental Well-being Support

Response Generation for Cognitive Behavioral Therapy with Large Language Models: Comparative Study with Socratic Questioning

Script-Strategy Aligned Generation: Aligning LLMs with Expert-Crafted Dialogue Scripts and Therapeutic Strategies for Psychotherapy