Interview of Emmanuelle Aboaf
summary
Live captioning and on-device transcription already enable participation but current systems misrecognize accented and deaf speech and require human correction; Microsoft PowerPoint and consumer apps demonstrably provide real‑time captions yet show frequent errors for French and nonstandard pronunciations. Proposes using AI/language models plus targeted voice datasets to raise accuracy, embedding transcripts on-room displays, and involving disabled users in product design — Microsoft PowerPoint live captions, accent data gap, nothing about us without us.
AI-written summary · about
related talks
48mProgramming for Accessibility
30mHear! An Accessibility Tale from a Hearing-Impaired Senior Software Engineer
25mKeynote Sports Panel Discussion
22mNothing for Us, Without Us; Breaking Unconscious Bias in Building Products
10m#BuildPropulsionLab at #DevoxxUK: @jbaruch with @antetmc2
49mThe impact of AI on people with disabilities
40mMicrosoft AI for Inclusion and Innovation
42mAccessibility is for everyone, including you!
