Events
What AI Can't Hear: Human Judgment in Accessible Audio Design
Centre for Multimodal AIFor decades, accessibility was treated as an afterthought, a ramp added after the building was already built. Today, AI risks becoming the new missing ramp: assumed to solve access problems that, without human oversight, it quietly creates. In her second edition of Inclusive Design for a Digital World, Reginé Gilbert argues that AI is not judgment, and that automated tools test against a model user that most real users simply aren't.
This talk examines where AI audio tools succeed, where they fail, and what gets missed when human decision-making is removed from the process. A passing automated check is not the finish line. The most dangerous answer in accessibility is the one that sounds right. This session asks: what does it mean to truly design for real people when the tools we rely on can't yet hear the difference?
Bio: Reginé Gilbert is a Systems & Change Strategist, Associate Professor at NYU Tandon School of Engineering, and Affiliate Researcher at Queen Mary University of London. She is the author of Inclusive Design for a Digital World (Apress, 2025) and Human Spatial Computing (Oxford University Press, 2026), and co-editor of Digital Accessibility Ethics (Taylor & Francis, 2026). Her work sits at the intersection of human judgment, emerging technology, and organizational change.
Updated by: Iran Roman

