To prevent false-positives, who knows if they have a "review" team that takes a tiny little peek at my naked son. Do you know for sure that no such system is in place or ever will be in place? What if they do get hundreds of thousands of false positives in the future? How would they improve their system if not by reviewing the existing system with real data?
They are not looking at your photos at all, they are comparing hashes of them to hashes of KNOWN predatory material. Your family photos will not be in that database.
I remember the Twitter account takeover hack a year ago, which was possible only because customer support had some 2FA-bypass: https://en.wikipedia.org/wiki/2020_Twitter_account_hijacking
To prevent false-positives, who knows if they have a "review" team that takes a tiny little peek at my naked son. Do you know for sure that no such system is in place or ever will be in place? What if they do get hundreds of thousands of false positives in the future? How would they improve their system if not by reviewing the existing system with real data?