When I edit the subtitles that are automatically generated by YouTube, I find that some videos say “you” or “so you” or “so very you” or “foreign” or “do foreign” at the end. It strikes me as odd.

Is this some kind of artefact? Does this happen to other people’s subtitles too?

Does it mean that the YouTube process does not know how to interpret the start of music? But it’s always the same music, so why would it sometimes hear “foreign” and sometimes hear “so you”?