International Journal of Scientific Research and Engineering Development

Growing dependence on digital devices and online work environments has made sustained cognitive effort a near-constant feature of student and professional life. Screen fatigue, irregular schedules, and persistent multitasking gradually wear down mental resilience, often long before users notice the signs. When cognitive burnout is left unaddressed, it erodes concentration, disrupts emotional regulation, and impairs judgment, outcomes with serious consequences in both academic and occupational settings. Conventional screening tools such as questionnaires and clinical interviews capture burnout only at fixed points in time and are heavily shaped by self-perception bias. This paper presents a Cognitive Burnout Early Warning System that sidesteps those limitations by drawing on passively collected interaction data together with Explainable Artificial Intelligence (XAI) to flag risk continuously and transparently. The framework tracks behavioural signals including active session length, idle intervals, keystroke cadence, interaction rate, after-hours device use, and sustained-activity streaks. Three complementary learners—Random Forest for risk classification, Logistic Regression as an interpretable baseline, and Isolation Forest for anomaly detection—work in concert to assign burnout risk levels ranging from low to high. SHAP values then decompose each prediction into per-feature contributions, giving users and clinicians a clear view of what drove the alert. A conversational assistant sits at the interface layer, translating model outputs into plain-language guidance: pacing suggestions, break reminders, and evidence-informed coping strategies tailored to the user’s current risk profile. The combined result is a system that detects deterioration early, explains its reasoning, and nudges users toward healthier work habits before burnout takes hold.

Dr. Prabha R, Nethravathi D, Neethu Chauhan, Sandra K P, H Anjana,"Cognitive Burnout Early Warning System Using Explainable AI and Conversational Assistant" International Journal of Scientific Research and Engineering Development, V9(3): Page(686-690) May-June 2026. ISSN: 2581-7175. www.ijsred.com. Published by Scientific and Academic Research Publishing.