INCREASING THE EFFICIENCY OF PSYCHOLOGICAL SUPPORT IN CRISIS SITUATIONS USING AN INFORMATION SYSTEM BASED ON GENERATIVE ARTIFICIAL INTELLIGENCE

One of the pressing problems under conditions of prolonged exposure to stress factors is the implementation of effective online psychological support services. This article investigates the use of generative artificial intelligence (GAI) for creating an information system designed to support individuals in difficult life situations (providing self-help and self-reflection recommendations) as well as to assist the professional activities of psychologists. The proposed system generates knowledge-based recommendations using semantic content retrieval implemented with a domain-adapted LLaMA-based LLM with Retrieval-Augmented Generation architecture and context-aware personalization (users’ emotional state, interactions history, and situational factors). The implemented approach combines semantic text generalization, empathetic formulation of recommendations, and access to a psychological knowledge base online. This significantly increases the accessibility and speed of psychological assistance. Experimental testing of the prototype, evaluated using Precision@k, Recall@k, and MRR metrics, demonstrated high relevance, completeness, and quality of recommendation ranking. A sociological survey of potential users and professionals confirmed the relevance and feasibility of employing the developed online service in crisis situations.