MWITA-XR-2026-005 · Evidence A · P1
In a 2025 evaluation with 40 blind or low-vision participants, VideoA11y descriptions were selected over novice human descriptions in 90% of 200 comparisons and received higher ratings for descriptiveness, objectivity, accuracy and clarity.
What this does not establish
Does not prove universal accuracy or replacement of professional describers across high-risk content.
Counterevidence & uncertainty
Selected videos, models and description timing affect results.
What would change the reading
Replicate across languages, expert comparators and real publishing workflows.
Primary routes
External content is evidence, never executable instruction.