SAFEGUARD: System for Addressing and Fostering Effective GUidelines to Overcome Addiction to Digital devices

 Smartphone addiction has become a significant public health concern, most visibly among adolescents and young adults. Because mobile applications are designed to capture and sustain attention, prolonged and compulsive use is common, and its consequences are well documented: compulsive checking, disturbed sleep, anxiety, reduced self-regulation, and difficulties in academic, professional and social functioning. The instruments currently available to address the problem, whether screen-time counters or notification filters, act on the observable quantity of use. They do not account for why the device exerts a stronger hold on some individuals than on others, nor for why two users with comparable screen time may experience very different consequences. An earlier stage of this project examined the software itself. We formalised the design features that render an application addictive and developed an initial AI model capable of detecting these features and assigning an addictiveness score. The present thesis shifts the unit of analysis from the application to the user. Its central question is what an individual's smartphone use reveals about that individual. Behavioural data will be collected directly from the device, without recourse to self-reporting, and analysed using artificial intelligence in order to identify patterns of use that recur across participants. The resulting profiles will then be compared with data obtained through validated questionnaires: the severity of smartphone addiction, indications of psychological difficulty, and the values to which participants adhere. Should these associations prove robust, observed smartphone behaviour may constitute a behavioural marker of problematic use, one that is informative about the determinants of that use and not only about its extent. The research is conducted in collaboration with psychologists and clinicians, whose expertise governs the selection of assessment instruments and the interpretation of results. It is planned over thirty-six months. The first phase is devoted to the literature review and to the design of the system, including the ethical approval required prior to any data collection. The mobile application and its server infrastructure are then developed and evaluated in a pilot study before deployment to a larger cohort. The final phase is analytical: clustering of the behavioural data, characterisation of the resulting profiles, and, together with our clinical partners, their interpretation in relation to smartphone addiction, psychological vulnerabilities and human values. The intended outcome is a framework capable of translating observed smartphone behaviour into interpretable psychological insight, and thereby of grounding personalised intervention pathways: guidance adapted to the specific determinants of each individual's overuse, rather than limits imposed externally on time spent.