Overuse of digital technologies and constant distractions have become critical challenges in today’s technology-driven environments. While Digital Self-Control Tools (DSCTs) aim to address these issues, many rely on static interventions that lack personalization and adaptability. In this work, we present Smart Adaptive Interfaces, a browser-based tool that modifies web content in real-time using Large Language Models (LLMs). The system contains three interaction modes, which allow users to suppress, obscure, or highlight web elements based on AI-identified distractions or dark patterns. We conducted a mixed-methods user study involving 18 participants to evaluate how these adaptive modes influence user experience and perceived control on two researcher-selected and one participant-selected website.
‹ All reading
AI-driven adaptive website content for digital well-being
Smart Adaptive Interfaces is presented, a browser-based tool that modifies web content in real-time using Large Language Models (LLMs), which contains three interaction modes, which allow users to suppress, obscure, or highlight web elements based on AI-identified distractions or dark patterns.