A Study on Universal Design Evaluation Methods in an AI-Powered Environment: Based on Experiments in Japan and Sweden Focusing on User Scanning Behaviour

This study examines how optimising information presentation in personalised AI improves usability while also introducing confusion and anxiety due to differences in users' cognitive characteristics and cultural backgrounds.This study aims to re-examine methods for evaluating universal design (UD) common to diverse users by focusing on "scanning behaviour," which is how users search, select, and interpret information. To explore this issue, this study is based on a longitudinal analysis of UD evaluation reports by Japanese students and a purchasing behaviour experiment comparing students in Japan and Sweden. The results showed that, in addition to conventional elements such as visibility, the way users rely on visual cues changes depending on differences in information access conditions and experiences. These results indicate the need to view user diversity not as attributes, but as a process of information search and cognition, and provide a new perspective on universal design evaluation in the age of AI.