Early Risk Identification and Support System for Mental Health Using Artificial Intelligence
Yousef Basuni1, Emad Abaalkhail2, Abdullah F. Basiouni3
1Yousef Basuni, Student, Education Sector, Royal Commission for Yanbu project, Saudi Arabia.
2Emad Abaalkhail, Researcher, King Abdulaziz City for Science and Technology (KACST), Saudi Arabia.
3Prof. Abdullah F. Basiouni, Yanbu Industrial College, Saudi Arabia.
Manuscript received on 18 December 2025 | First Revised Manuscript received on 24 December 2025 | Second Revised Manuscript received on 04 January 2026 | Manuscript Accepted on 15 January 2026 | Manuscript published on 30 January 2026 | PP: 1-5 | Volume-15 Issue-6, January 2026 | Retrieval Number: 100.1/ijsce.F370815060126 | DOI: 10.35940/ijsce.F3708.15060126
Open Access | Editorial and Publishing Policies | Cite | Zenodo | OJS | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open-access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: The burden of mental illnesses, especially depression and anxiety, is high in the world, and in most cases, it results in severe losses of quality-adjusted life years. This paper describes advancements and initial estimates for an artificial intelligence (AI) system expected to diagnose mental health risks early and provide individual-level support. The technique impacts Natural Language Processing (NLP) and emotion analysis to identify emotional structures in user-posted text, such as daily diaries and mood journals. An emotional tone Bidirectional Encoder Representations from Transformers (BERT) model is fine-tuned, and the system suggests self-care options (e.g., mindfulness exercises, breathing) in response to the context, towards an adaptive recommendation engine. One notable aspect is a user friendly visual dashboard that enables users to monitor their mood patterns over time. More importantly, the system is entirely offline, and the user’s privacy is guaranteed, as all data is processed locally on the machine. The data simulation tests the system’s functionality for sentiment classification and recommendation delivery. The results indicate that this platform may be a promising, ethics-driven, proactive mental health support tool and may be applied in educational, workplace, and personal contexts. The next phase of work will be long-term real-world validation and efficacy studies.
Keywords: Mental Health, Artificial Intelligence, Natural Language Processing, Sentiment Analysis, Early Intervention, Digital Health, BERT Model.
Scope of the Article: Artificial Intelligence