Hacker News new | ask | show | jobs
by ncw96 1774 days ago
A summary of the photo scanning system:

- Only applies to photos uploaded to iCloud

- Matching against a known set of CSAM (Child Sexual Abuse Material) hashes occurs on-device (as opposed to the on-server matching done by many other providers)

- Multiple matches (unspecified threshold) are required to trigger a manual review of matched photos and potential account suspension

3 comments

You should reread the first section of TFA, titled "Communication safety in Messages." This goes beyond the scope of CSAM: they're scanning all Messages photos sent to or from a minor's account for any possible nudity, not just CSAM hash-matching.
It sure seems like this is two different techniques/technologies at work. There is a CSAM deterctor using a known database and then, it looks like, there is a separate model in messages for detecting pornographic content without a known database.
It doesn't sound like it can "detect pornographic content" since the difference between pornography and nudity is not going to be reducible to a coefficient matrix.
Yeah but that just blurs the image and notifies the parents. Pretty low impact for false positives.
Sorry for the confusion — I was referring to just the CSAM hash feature that uploads results to iCloud.

There is also scanning for nudity in the Messages app, but those scans happen on-device and the photos stay on-device even if nudity is detected.

Apple has been scanning iCloud for child abuse since (at least) Jan 2020 [1] [2]

[1] https://www.telegraph.co.uk/technology/2020/01/08/apple-scan... [2] https://web.archive.org/web/20200110193302/https://www.apple...

Read the article again. The on device scanning makes hashes on the local device