Resource article
Auto-generated transcripts vs ADA-compliant closed captions
Why machine-generated text can help draft workflows but does not automatically satisfy accessibility obligations.
Direct answer
What this means in practice
WCAG 2.1 SC 1.2.2 requires captions for prerecorded synchronized media, with limited exceptions.[1]
Caption requirements include spoken dialogue and meaningful non-speech sounds, not just word-for-word transcript text.[1]
YouTube documents that automatic captions can misrepresent speech and should be reviewed and edited.[2]
A compliant workflow typically includes human quality checks for accuracy, timing, speaker context, and formatting.[1][2]
FAQ
Are transcripts and captions the same thing?
No. Captions are synchronized to media playback and include non-speech audio context. A plain transcript alone does not provide that synchronized experience.[1]
Can we start from automatic captions?
Yes, but YouTube itself recommends reviewing and correcting auto-captions because machine output may be inaccurate.[2]
Annotated sources
- [1] W3C WCAG 2.1 Understanding SC 1.2.2 (Captions, prerecorded)
W3C WAI | WCAG 2.1 guidance
Defines what captions must include and why synchronized captioning is required for accessibility.
- [2] YouTube Help: Use automatic captioning
YouTube Help | Help article
States auto-captions may misrepresent speech and should be reviewed/edited.
This article is informational and not legal advice. Organizations should consult counsel for legal determinations.