YouTube is introducing machine-generated automatic captioning to YouTube. The captions can also be translated. This potentially might have considerable implications for the hearing-impaired and language translation, albeit it will need further work in terms of reliability. Automatic captions will be generated using Google’s automated speech recognition (ASR) technology and the same voice recognition algorithms used in Google Voice. Additionally, auto-timing is being introduced. If you provide all the words in the video, Google will automatically time the captioning for you.
Google put together a video on how to access the automatic captioning and auto-timing features:

Source: Blog.searchenginewatch.com