Mental health challenges and suicidal ideation are critical public health concerns that require timely intervention and accessible support systems. This project, “AI-Based Suicide Risk Detection and Emotional Support Chatbot,” presents a local, privacy-preserving conversational agent designed to provide empathetic emotional support and identify high-risk suicidal language in real time. The system integrates a pre-trained machine learning classifier using TF-IDF and Logistic Regression for suicide risk prediction, along with a free, offline Large Language Model (LLM) powered by GPT4All for supportive conversation. The chatbot dynamically analyzes user input to assess suicide risk levels and responds accordingly. In high-risk situations, it prioritizes safety by providing immediate helpline resources, while in low-risk interactions, it offers compassionate conversation, emotional validation, and well-being recommendations such as music, movies, breathing exercises, and motivational quotes. The system emphasizes user privacy by operating entirely on local resources without cloud dependency. This project aims to support early emotional intervention, reduce stigma around mental health discussions, and complement existing mental health care systems.
sudharsan Sudharshan (Mon,) studied this question.