| KEYWORD |
Unhooked From the Smartphone: Leveraging Large Language Models to Detect Attention Traps
Parole chiave AI, ATTENTION-CAPTURE DESIGN PATTERNS, DIGITAL WELLBEING, LARGE LANGUAGE MODELS, MOBILE APP DEVELOPEMENT
Riferimenti ALBERTO MONGE ROFFARELLO
Gruppi di ricerca DAUIN - GR-10 - Intelligent and Interactive Systems - ELITE
Tipo tesi EXPERIMENTAL, RESEARCH
Descrizione As digital platforms increasingly rely on attention-capture design patterns, such as infinite scrolling or content autoplay, users often find themselves spending more time than intended on apps. These patterns exploit cognitive biases, drawing users into prolonged engagement and undermining their digital autonomy. While various solutions exist to help users manage screen time, few offer proactive, real-time detection of these manipulative design elements. Leveraging advancements in large language models (LLMs), we now have the opportunity to create intelligent systems that recognize these patterns and provide users with timely interventions, promoting healthier digital habits.
This thesis explores the development of a mobile application designed to enhance digital well-being by leveraging large language models (LLMs) to detect attention-capture design patterns in real time. The research aims to identify moments when users are subject to manipulative design elements that undermine their control over time spent on digital platforms. By detecting these patterns, the app will provide interventions to help users regain agency and manage their digital habits more consciously. Operatively, the app will be ideated and prototyped for the Android operating system, and assessed through a field experiment measuring quantitative and qualitative impacts on users' smartphone habits. The outcome will contribute to digital well-being research and propose novel design strategies that align with ethical, user-centered principles in mobile app design.
If satisfactory, the result of the thesis will be released as an open-source project.
Conoscenze richieste Good programming skills and understanding of Android app development are required.
Scadenza validita proposta 14/10/2025
PROPONI LA TUA CANDIDATURA