There’s already an existing feature suggestion discussing an option for automatically generated captions/subtitles. I suppose that refers to a literal transcription as well as an option to include translated versions. In that case I suggest you comment/vote in that thread in order to help us avoid distracting or cannibalizing votes (e.g. focusing on the feature / use case, and not necessarily the particular technology to be used to deliver such a feature).
Unless I’m missing an aspect of your suggestion I’ll close it as a duplicate.
