Better AI with Designed Friction: Theories, Applications and Research Agenda

Résumé fourni par la source

The design of AI systems prioritises frictionless interaction, often at the cost of human judgement and accountability. This contribution discusses designed friction—deliberate inefficiencies designed to foster engagement, critical reasoning, and user agency. Drawing from friction-in-design and responsible AI, we explore key mechanisms such as cognitive forcing functions, seamful design, and programmed inefficiencies. We propose a research agenda for Hybrid Human-AI systems that balance usability and trust through frictional AI, aiming to provide a stepping stone for the future development of a dedicated research community.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.