An AI based privacy oriented approach to administer Psychological Projective Test

In recent years, there has been a sharp rise in the demand for mental health services, with an increasing number of individuals seeking assistance for psychological and emotional challenges. This escalating need for mental health support has placed a significant strain on healthcare professionals, necessitating innovative solutions to manage the growing workload effectively. This research introduces a pioneering approach that leverages Large Language Models (LLMs) to automate psychological triage, offering a potential solution to reduce the burden on mental health professionals. The primary objective of this study is to streamline the psychological assessment process while ensuring patient anonymity and data privacy. The proposed framework centers on the administration of the Thematic Apper- ception Test (TAT) on edge devices, followed by facial emotion recognition and text transcription performed by existing state-of-the-art Neural Networks. The collected data is securely encrypted and transmitted to an LLM for comprehensive analysis. The result is the generation of an extensive PDF report, which can be shared with specialists to gain an initial understanding of the patient’s profile. In a world where the demand for mental health services continues to grow, this research presents a novel perspective on psychological triage, aiming to strike a balance between efficiency, patient privacy, and in-depth analysis. This work includes an introduction to the problem statement, a theoretical background of the tools used to realize the framework, a description of the methodology adopted, the results obtained as well as some challenges and possible future development.