Apple isn't law enforcement. Of course it's not their place to be stopping the sexual exploitation of children. They're selling storage devices. Has anyone ever suggested that closet and basement makers should install covert scanners to make sure no one is hiding kidnapping victims, that automakers should put scanners in the trunks to make sure nobody is hauling off murder victims to bury, that makers of filing cabinets should install scanners to ensure you're not storing photos of known exploited children in your filing cabinet? Why is this topic treated so differently when the devices become networked? We could just as easily tap every phone line in the world and zoom in on some region where a child was just reported kidnapped and flag all people who say their name over a phone line, but we don't do that, and no one in their right mind would find that acceptable. The police need probable cause and a warrant before they can install a listening device on your phone line. They don't pre-surveil everyone to flag potential criminals before they become known by other means. I don't understand why software services are held to such a lower privacy standard than every other form of storage and communication we have.

And to specifically answer the question of what they can do. When the police come with a warrant because they have good reason to suspect some customer X is storing and/or transmitting child porn, then transmit a covert over the air update to that customer's device and that customer's device only, scanning photos on-device before they get encrypted for iCloud backup. Don't push the surveillance update to every single customer on the planet before they're under any suspicion and this suspicion has been certified by a judge just in case.

>But if your position is that they shouldn't be looking at your data at all, then you have to go back to the first question and say "Apple should not be trying to stop the exploitation of children."

You're getting pretty far into "affirming the consequent" territory here.

That's fair. I guess a better answer would be "Apple should not be trying to stop the exploitation of children by looking at user's data."

But what else could they meaningfully do?

Donate to NCMEC. They are woefully underfunded...

I don't think donating to a private yet government funded, nontransparent and unaccountable organization is a good idea. In fact, I don't think that having such an organization is a good idea at all.

> But... "what about the children?" is a very important question to ask and shouldn't be used as a cliche to dismiss Apple's attempt to make the world a better place in this area.

I've never seen a single person ask this question in good faith. It never stops. Every single day we have a new crisis involving children and the solution is surveillance, more government control, less freedom, whatever.

Nothing but emotional manipulation. None of this will ever stop crime. None of this will ever undo the suffering already inflicted. All of this is mainly justified by the debatable notion that demand for CSAM causes child molestation, the actual crime authorities should be trying to prevent. Instead of doing that they use children in order to establish the surveillance tools needed for government oppression. Using children as a political weapon should be considered child abuse as well.

>They could do whatever ever other company is doing and just looking at all of your data, but Apple didn't want to do that.

Note that Apple doesn't do end-to-end encryption, there's little evidence they are even thinking about it, and there are other issues that would stand in their way if they had tried to. Under the current system, we get client-side scanning and server-side scanning whenever Apple feels like it. Right now, there's no privacy benefit other than the made-up one some Apple fans invented from thin air.

>So they came up with a system that they felt was a good compromise between privacy and trying to stop the spread of CSAM.

It's a terrible 'compromise'. In return for a bad precedent, they get a system that catches way less cases than the typical server-based ones. Unless Apple expands the system further to catch these cases too, in which point one should ask how much privacy is really kept.

Cloud scanning would be better, because there’s a clear way to opt out - don’t back up to the cloud. Scanning in device takes a hard technical restriction where they can’t scan files they don’t have access to and turns it into a policy restriction where they say they won’t scan files that are not being uploaded to iCloud.

How long do you expect that policy to realistically hold when some three-letter-agency comes knocking?

If they don‘t upload there is no information anywhere, but on your device. The system very much depends on the cloud, because the keys of known photos aren‘t stored locally. Photos are already scanned locally for the keyword search and faces,and that could easily be used to search for anything, but so far no one had an issue with that.

If you say, 'Yes, apple should be doing something, but not this." I'd love to hear your ideas about how Apple can do more to stop CSAM while better protecting your privacy.

Detecting CSAM on a phone is not stopping exploitation of children. Just as detecting images of violence isn't stopping violence. Violence and abuse have existed before phones and cloud storage, and will continue to be ever after.

And Apple cannot stop exploitation of children. Anything they try is trivial to circumvent, a massive breach of privacy, or both.