Alt-text Isn't Enough - MDLS 2026

Starting with the wide availability of frontier AI models, accessibility conversations in scholarly communication have converged toward a single fix: add alt-text and call it done. Machine generated alt-text can be a significant improvement, but leaves so much rich information out of reach while STEM researchers, students, and the public are left grappling with figures and data that could benefit from a holistic approach to accessibility. Drawing on the Web Content Accessibility Guidelines (WCAG) and the compliance pressure created by ADA Title II and the European Accessibility Act (EAA), this session will walk through the accessibility gaps that alt-text alone can’t close, and why closing them improves usability for nearly everyone in the room, not just those who identify as disabled. Accessibility is the foundation of a truly open scientific record. The talk will also draw a direct parallel to the FAIR data movement. FAIR succeeded because campuses built infrastructure — library services, repositories, mentorship — around making data machine-accessible. Accessibility for human readers could benefit from this approach. I’ll make the case that data librarians are uniquely positioned to close that gap, since many already sit at the intersection of data services, instruction, and publishing support.