Design And Development Of A Psychological Assessment Tool Based On LLM Agents

Abstract

This work aims to design and develop a psychological assessment system based on language models. by leveraging these latter's with advanced integration and utilization, The goal is to create an intelligent multi-agent system that can assist professionals psychotherapists and patients, This agentic tool is not only technically robust but also ethically aligned, mitigating risks such as hallucination and bias through prompt constraints and evidence-grounded response generation. utilizing techniques such as instruction tuning, domain-adaptive fine-tuning, and quantization, the platform achieves scalability and adaptability in resource-limited environments. By integrating advanced AI methodologies with mental health expertise, the system provides emotionally sensitive support to users while automating and enhancing documentation and decision-making tasks for clinicians. This work represents a significant step toward ethically grounded, context-aware, and clinically viable AI in mental healthcare applications. as it demonstrates the effectiveness of the medical assistant in enhancing healthcare delivery. Through