Text and web accessibility for neurodiverse populations, including individuals with autism, is a widely recognized concern. However, many existing guidelines for these adaptations lack empirical evaluation. Additionally, the challenges in diagnosing autism in adulthood lead to unmet support needs for many adults on the spectrum. In this keynote talk, we present experiments utilizing eye-tracking data to address these two seemingly unrelated needs simultaneously: enhancing text and web accessibility and facilitating autism detection in adulthood. Gaze data from adults both with and without Autism Spectrum Disorder (ASD) were recorded during text reading and web information retrieval tasks. Comparing the two groups, we examined their efficiency and accuracy in answering reading comprehension questions and finding target information within web pages. The results suggest that the ASD group invests more cognitive effort to achieve the same accuracy results as their neurotypical counterparts, and that this is likely due to a sequential approach to information searching. Next, we hypothesize that the differences captured by the eye-tracking data can be used to train an autism detection classifier. Several experiments show a consistent classification accuracy of around 75%. We discuss the implications of the produced evidence of visual processing differences in adults when reading and searching for information in web pages for advancing accessibility research and autism screening.
