Mobile smartphones now serve as essential tools across education, professional duties, personal communication, and daily administration. Despite their utility, habituated checking behaviors, prolonged sessions, and poorly timed notifications frequently disrupt personal productivity and cognitive focus. Standard operating system utilities supply static screen time metrics and app limits, yet struggle to account for user intention or mental context. Recent research in human-computer interaction and mobile computing explores personalized, machine learning-driven interventions designed to adapt to individual usage routines. This survey reviews contemporary developments across five key domains: device telemetry extraction, self-regulation interfaces, micro-nudging mechanics, context-aware notification scheduling, and human-centred artificial intelligence architectures. We synthesize existing empirical studies across performance metrics, data collection techniques, personalization depth, and methodological constraints. The findings indicate that raw usage duration is a weak signal for psychological strain, while intervention success hinges on environmental context and individual motivation. Key challenges remain regarding data privacy, model interpretability, and long-term user agency. Finally, we present a modular architectural direction combining local telemetry processing, explainable baseline modeling, uncertainty estimation, and user-governed intervention controls.

Manasvi Patil, Pranav Desai, Aniket Malik, Adinath Kashte, Prof. Nilima Matre, "A Survey of AI-Driven Digital Wellbeing Systems: Smartphone Behaviour Analysis, Personalization, and Adaptive Interventions" International Journal of Computer Science and Engineering Development, V1(4): Page(195-199) September - October 2026. ISSN: 3139-0862. www.ijcsed.com. Published by Scientific and Academic Research Publishing.