Understanding 1.2.4 Captions (Live)
"Enable people who are deaf or hard of hearing to experience real-time presentations. Captions provide the part of the content available via the audio track."
This criterion applies specifically to live digital presentations like a webinar or a virtual conference. The purpose is to ensure that all listeners, especially people who are deaf or hard of hearing, are able to access information in real-time when watching or listening to live media.
To do so, you must provide synchronized, real-time captions of all audio content, with no more than a broadcast delay. This can either be delivered as closed or open (always visible) captions, or via a synchronized text stream.
Human vs. AI-generated captions
Technology has come a long way since we first published this guide in 2019. Today, almost all video conferencing and streaming platforms provide closed captioning options, many with auto-captioning services at no additional expense. Similarly, there are far more human captioning service providers available. Which is best for achieving 1.2.4 compliance?
To be 1.2.4 compliant, accuracy and speed are paramount. While AI-generated captions happen in real-time they are almost always plagued by errors and the experience is far less accurate. Human-generated captions on the other hand tend to be more accurate, but delivered with a greater delay. There is also a slight risk of entire blocks of content getting missed with human captioning.
Recommended Solutions
The ideal solution is actually a combination of both human and AI-generated captions. There are service providers who are leveraging AI-driven solutions that are manually moderated and modified by a human, with a slight delay.
