Hacker News new | ask | show | jobs
by read_if_gay_ 1777 days ago
If this technology is that easily circumvented then why is there an expectation that it will be effective at all?
3 comments

The CASM scanning happens on device, right? At least so we’ve heard.

My sense is Apple is trying to keep CASM off their servers. Scanning phones before it gets there was their solution to what I assume is a government demand/ultimatum. “Do this or we repatriate your foreign entity taxes” or some other shit.

I too feel that Apple just caved and eroded trust that took decades to build up. The only way this gets sorted is the “screeching minority” continues to screech and brings other in. Notify state attorneys general, FTC, etc. will that do anything? Who know? My bet is that it’s the DOJ behind all of this.

Hopefully the plaintiff bar which are already preparing class action lawsuits will find a way to get documents in discovery that allude to government coercion. But then again I’m sure there would be a clever way those are not produced under some “national security” bullbaiting reason.

All we can do is try, and keep the pressure on.

> My sense is Apple is trying to keep CASM off their servers

It could (maybe) also be a prelude to enabling E2E encryption for everything in iCloud.

That's my theory as well. I see all those people selling their iPhones that will instead use Android, upload all their photos in Google Photos, and Google will happily share all those those same photos. Yes, yes, yes, I know, one is done on device, the other one is done in the cloud, for me that is pretty much 2 sides of the same coin.
As I understand it (and I've not spent too long on this, just picking at various articles) - there are two separate things at play here.

Firstly - CASM scanning is done via fingerprinting - the image is fingerprinted on device and when uploaded to iCloud that fingerprint is compared with the "dodgy images" fingerprints and an alert raised if a threshold of matches is reached (what's the threshold and with whom?)

Secondly - there is on-device AI image recognition - when you send an image to someone else (via iMessage or the share sheet) it is checked for nudity and if the iCloud account in question is registered to a 13-year old or younger, their parents are alerted.

In both cases the fingerprinting/scanning is on-device and is triggered by the images leaving the device.

> the image is fingerprinted on device and when uploaded to iCloud that fingerprint is compared with the "dodgy images" fingerprints and an alert raised if a threshold of matches is reached (what's the threshold and with whom?)

Nope. The comparison is done on the device and the threshold is set there as well.

I am not sure how alarmed I am yet at this whole affair but I do know that maybe 50% of posts I read about this have glaringly incorrect information which definitely dampers my alarmism.

> Nope. The comparison is done on the device and the threshold is set there as well.

As I understand it the fingerprinting and comparison is done on device, but it only happens as part of the upload-to-iCloud process. So the grandparent's phrasing isn't unreasonable.

It isn’t being circumvented. It is intended to work on photos that are uploaded to iCloud. If you don’t use that (via turning it off or via selling everything Apple and switching to Linux) then you aren’t using it.
No. The intent is not that it works on iCloud. The intent is catching pedos.

This is obviously not effective given that you can get around it that easily if you want to. Coincidentally though, it will be totally effective at surveilling the 99.999% that are normal users and won’t go out of their way to disable iCloud. The whole CP thing is such an obvious farce.

The theory would be that many people are stupid. Of course we only know about criminals who get caught and that tends to be because they made a mistake, so it looks like most criminals are stupid.