A Structured and Safety-Aware Conversational Framework Using Large Language Models and Retrieval-Augmented Generation for Mental Health Support
International Journal of Advanced Research in Science Communication and Technology · 10.05.2026 · 770
DOI: 10.48175/ijarsct-34979 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Advanced Research in Science Communication and Technology, online erschienen am 10.05.2026. Verfasst von Mr. Varad Deokar und 3 weiteren Autor:innen. Die Arbeit nennt 13 Literaturangaben.
Abstract (Originalsprache)
The growing demand for accessible mental health support has led to the development of AI-based conversational systems; however, many existing solutions either lack structure or fail to ensure safety and consistency. This paper presents a structured and safety-aware conversational framework that integrates Retrieval-Augmented Generation (RAG) with Large Language Models (LLMs) in a controlled, multi-stage pipeline. The proposed system incorporates dedicated modules for crisis detection, assessment, cognitive reflection, and intervention, coordinated through an intent-driven orchestration layer. Unlike conventional chatbots, the system emphasises deterministic dialogue flow and psychological structuring using established frameworks. Evaluation using simulated conversational scenarios demonstrates that the framework produces more coherent, context-aware, and safe responses compared to unstructured approaches. The results highlight the importance of architecture-driven design in building reliable and responsible AI systems for mental health support
Zitieren
APA
Deokar, M. V., More, M. A., Parmar, M. M., & Gangrade, P. K. (2026). A Structured and Safety-Aware Conversational Framework Using Large Language Models and Retrieval-Augmented Generation for Mental Health Support. International Journal of Advanced Research in Science Communication and Technology, 770. https://doi.org/10.48175/ijarsct-34979
Vancouver
Deokar MV, More MA, Parmar MM, Gangrade PK. A Structured and Safety-Aware Conversational Framework Using Large Language Models and Retrieval-Augmented Generation for Mental Health Support. International Journal of Advanced Research in Science Communication and Technology. 2026:770. doi:10.48175/ijarsct-34979
BibTeX
@article{Deokar2026,
author = {Deokar, Mr. Varad and More, Mr. Atharva and Parmar, Mr. Manav and Gangrade, Prof. Kopal},
title = {A Structured and Safety-Aware Conversational Framework Using Large Language Models and Retrieval-Augmented Generation for Mental Health Support},
journal = {International Journal of Advanced Research in Science Communication and Technology},
year = {2026},
pages = {770},
doi = {10.48175/ijarsct-34979}
}
EUVIMED