Best guess would be using the database that sites like tumblr and twitter use to automatically identify known CSAM. I’m not sure how it works, I just know that sites use a database to quickly shut down any uploads
Comment on I just developed and deployed the first real-time protection for lemmy against CSAM!
ram@feddit.nl 1 year ago
I’m curious. How do you train such AI without being raided by the authorities?
Callie@pawb.social 1 year ago
Deepus@lemm.ee 1 year ago
Offload all work to an anonymous VPS provider possibly? I dunno just spit balling.
god@sh.itjust.works 1 year ago
You don’t tell them.