EXAD: Framework for Designing Explanations in Adaptive User Interfaces

Adaptive user interfaces that personalize their behavior based on user patterns are becoming increasingly common, yet they often fail to communicate their adaptations to users effectively. This paper introduces EXAD, a framework for designing explanations in adaptive interfaces that considers adaptation type, explanation timing, and user characteristics. Through a mixed-methods study with 42 participants, we investigated how explanation approaches affect trust, understanding, and efficiency across content, layout, and interaction adaptations. Our findings demonstrate that explanation effectiveness varies significantly by adaptation type: Proactive explanations build the highest trust for layout adaptations, while progressive explanations offer the best balance between transparency and efficiency. We extend the EXAD framework to B5G (Beyond Fifth Generation) network systems, enabling context-aware explanations that enhance trust and transparency. Our results show that carefully designed explanation strategies improve user understanding and operational efficiency, and can be translatable in dynamic network environments such as resource allocation based on user patterns in B5G networks. This research bridges theoretical work in explainable AI with practical implementation challenges in adaptive interfaces.