This is entirely the fault of the IWF and Microsoft, who create “exclusive” proprietary CSAM prevention software and then only license big tech companies to use it.
This is entirely the fault of the IWF and Microsoft, who create “exclusive” proprietary CSAM prevention software and then only license big tech companies to use it.
Publishing a list of hashes would make it trivial for abusers to know when their images are being flagged. It would be better to get M$ to do the scanning work themselves
Bingo. It would also make it trivial to alter images just enough so that it wouldn’t match the hash, and then they can post shit that would need to be manually flagged and removed.
I already see things like this with pirated media; pirates will include extraneous material bundled with the target media so that it’s not automatically flagged and removed.