Apple's "Perceptual hash" which they said was "unreadable" has been extracted and collisions generated

posted 3 years ago by the-new-style 3 years ago by the-new-style +37 / -0

Code for extraction

https://github.com/AsuharietYgvar/AppleNeuralHash2ONNX

Working collision

https://github.com/AsuharietYgvar/AppleNeuralHash2ONNX/issues/1

Apple's claim

https://www.apple.com/child-safety/

Apple’s method of detecting known CSAM is designed with user privacy in mind. Instead of scanning images in the cloud, the system performs on-device matching using a database of known CSAM image hashes provided by NCMEC and other child safety organizations. Apple further transforms this database into an unreadable set of hashes that is securely stored on users’ devices.

And not only that, according to this reddit post

Believe it or not, [the NeuralHash algorithm for on-device CSAM detection] already exists as early as iOS 14.3, hidden under obfuscated class names.

10 comments

10 comments share save hide report block hide replies

Comments (10)

sorted by:

▲ 2 ▼

– Mad_King_Kalak 2 points 3 years ago +2 / -0

Thanks for the explination, and I surmise that you're correct about memes saved on your phone.

My apple is never getting synced again, and I will be getting a Gab phone when those come out. Alas, I lose years of text messages on the Apple and have to physically write my contacts down.

permalink parent save report block reply