The Americans With Disabilities Act Was Passed in 1990. Online Video Did Not Have to Follow It for Another Two Decades. Here Is Where Things Stand Now.
When President George H.W. Bush signed the Americans with Disabilities Act in July 1990, no one considered the implications for internet video. The World Wide Web did not exist in public form yet. YouTube was fifteen years away. The ADA was designed around physical spaces, employment, and telecommunications, not streaming media. The closed captioning requirements that existed applied to broadcast television under separate legislation, and they were enforced inconsistently for years.
The gap between the ADA's 1990 passage and meaningful online video accessibility compliance is a thirty-year story of technology outpacing law, advocacy organizations filing lawsuit after lawsuit, and institutions discovering that videos they had recorded and posted for years were legally problematic. The Department of Justice issued guidance in 2022 explicitly stating that Title III of the ADA applies to websites and web content, including videos. By then, universities, hospitals, news organizations, and businesses had already faced successful legal actions over uncaptioned video content.
The practical upshot is that any organization posting video content online needs to understand what accessibility requirements apply to that content and what producing compliant transcripts and captions actually requires. Transcription is the starting point for both.
The Legal Landscape: ADA, Section 508, and the CVAA
Three main legal frameworks govern video accessibility in the United States, and they apply to different categories of organizations.
The Americans with Disabilities Act, specifically Title III, covers places of public accommodation and commercial facilities. Courts have extended this to websites and web content through a series of landmark rulings. The National Association of the Deaf versus Netflix, settled in 2012, established that streaming video services must provide accurate captions. Similar actions have been brought against Hulu, Harvard University, MIT, and dozens of other organizations. The DOJ's 2022 guidance made explicit what courts had been finding implicitly: ADA Title III applies online.
Section 508 of the Rehabilitation Act applies specifically to federal agencies and organizations that receive federal funding. Section 508 requires that video content meet WCAG 2.0 Level AA guidelines, which include captions for pre-recorded video and transcripts sufficient to understand the content. This applies to the video content itself as well as to the player interface. A university receiving federal funding, which includes virtually every major US university, must comply with Section 508 for video content published on its platforms.
The 21st Century Communications and Video Accessibility Act, or CVAA, passed in 2010, closed a specific gap by extending closed captioning requirements to online video that had previously been broadcast on television. If a program aired on broadcast TV with captions and was then made available online, the online version must also be captioned. Together, these frameworks mean that caption and transcript requirements apply to a large portion of video content published online.
What Transcription Is and What It Is Not
A transcript is a complete text representation of what is spoken in a video. It captures all dialogue, narration, and other relevant audio content in plain text form. For simple video content with one speaker, a transcript may be sufficient for basic accessibility compliance in some contexts. For video content where the visual track conveys information that is not available in the audio, a descriptive transcript is required: one that describes relevant visual elements in addition to capturing the audio.
A caption is time-synchronized text that appears on the video screen as the audio plays. Captions serve a different function from transcripts. They allow deaf and hard-of-hearing viewers to follow the video in real time, with the text appearing at the same moment as the corresponding audio. Transcripts can be read independently from the video. Captions are consumed alongside the video.
WCAG 2.1 success criterion 1.2.2 requires captions for all pre-recorded audio content in synchronized media. At Level AA, which is the compliance level required by Section 508 and referenced by the DOJ ADA guidance, both captions and transcripts are relevant requirements depending on the content type. For most organizations, the practical approach is to produce a transcript first and then convert the transcript into time-synchronized captions. AI transcription makes the first step fast.
The Search Engine Dimension
Beyond legal compliance, transcription serves a significant search discoverability purpose. Search engines can index text. They cannot meaningfully index the spoken content of a video audio track. A one-hour video recording of a conference presentation may contain tens of thousands of words of expert content, case studies, and technical detail. Without a transcript, none of that content is accessible to search engines. The video exists as a title, a thumbnail, and whatever text appears in the surrounding page content.
A transcript attached to the video, embedded in the page as text, or provided as a separate document exposes all of that content to search indexing. Words that appeared only in the audio track become discoverable. A viewer searching for a specific term mentioned in the presentation can find the video. A researcher looking for a specific example discussed in the webinar can locate it.
This search value compounds across an organization's video library. Universities that have published hundreds of recorded lectures without transcripts have large bodies of educational content that search engines cannot access. Organizations that have posted webinars, product demonstrations, and training videos without transcripts have made accessibility and discoverability sacrifices simultaneously.
Accuracy and the Correction Step
AI transcription produces a draft. The draft is accurate for clear audio with standard speech. It requires correction for unclear audio, technical vocabulary, names, or accented speech that the model was not trained on specifically.
The correction step is faster than producing a transcript from scratch but is not optional for professional use. A transcript submitted for legal compliance, published on a university website, or used as a basis for captions should be reviewed for accuracy before use. Errors in transcription that result in inaccurate captions may not meet accessibility requirements.
For high-value content, the combination of AI transcription and human review is the standard workflow: AI produces the draft in minutes, a human reviews and corrects it in a fraction of the time it would take to transcribe manually, and the final transcript is accurate enough for professional use. The transcript output provides the foundation for whatever downstream use the content requires.
The Practical Compliance Test
For any organization with an existing library of video content, the compliance question is straightforward to assess. Take a sample of the most recently published videos. For each one, ask: does this video have accurate captions or a transcript accessible from the page where the video is published? If the answer is no for most of the sample, the organization has an accessibility compliance gap that represents both legal exposure and an audience that cannot access the content.
The path forward is to prioritize the highest-traffic and most recently published content, produce transcripts using AI transcription and human review, add captions to the video files or publish transcripts alongside the videos, and establish a process for new video content that includes transcription as a standard step before publication. The transcript is not an addition to the video. It is part of what makes the video complete.
Conclusion
The legal history of online video accessibility is a story of law catching up to technology, with enforcement actions accelerating the pace. Organizations that have published video content without transcripts or captions are not in an ambiguous situation. The DOJ guidance is clear, the case law is established, and the compliance path is well-defined.
ToolHQ's Transcribe Video tool processes video files securely on the server and deletes them immediately after transcription, generating a text transcript that can be reviewed, edited, and used as the foundation for captions, meeting notes, or published accessibility content.
Frequently Asked Questions
Does the ADA require transcripts for online videos?
Yes. The DOJ's 2022 guidance states ADA Title III applies to websites and web content including videos. Court cases against Netflix, Harvard, and MIT have established that uncaptioned online video violates the ADA.
What is the difference between a transcript and captions?
A transcript is a text document of the spoken content, readable independently. Captions are time-synchronized text that appears on the video while it plays. Both may be required depending on compliance framework and content type.
How accurate is AI video transcription?
AI transcription is accurate for clear audio with standard speech, producing a draft that is faster to review and correct than transcribing from scratch. Technical vocabulary, names, and accented speech may require correction.