It starts with comparing file hashes. I’m worried about where this goes next.

By that logic, how do you know it hasn’t already gone there, years ago?

And this scan is currently happening on the server. They’re just making it so that the scan will now happen at the time of upload as opposed to N seconds later.