(I think they've since delayed the rollout of that system indefinitely, after public outcry.)
(I think they've since delayed the rollout of that system indefinitely, after public outcry.)
Does it matter in practice? No. But it makes me feel very different about it. That's important too.
Sounds like you’re taking exception with the explicit parental notification, where a parent with a child enrolled in their iCloud “family” can request to be alerted when their minor child takes an action on the phone owned by the parent.
The EFF wrote an awful blog that deliberately confused the already confusing release from Apple. Your privacy is almost certainly weaker as a result, as various entities can use a subpoena or warrant to access your files.
And once you're scanning files with one flag set, nothing technologically prevents the scanning of files without that flag being set. And to quote myself from the Google Stadia brouhaha - "companies lie in PR statements" - so I have no reason to trust Apple's statement that they would never scan other files.
You need to read up on how the system worked, because they picked a design that made absolutely no sense if they wanted to do that. They’d have to redesign it to work in a different way if they wanted to do that.
The material that they scan for is illegal to possess. You should assume that anyone that accepts cleartext uploads of binary data is scanning for that and other material.
At the end of the day, you don’t have the juice to negotiate contractual protections, so corporate self-interest is really your only real protection.
Do you know what else is illegal to possess? Proof of abortions in some states. Proof of being gay in many countries. Winne the Pooh in at least one country.
> You should assume that anyone that accepts cleartext uploads of binary data is scanning for that and other material.
Again, since this seems to be lost every time, this isn't being checked post upload. That matters.
From a technical standpoint, at least you are protected from future policy changes if your files in the cloud are encrypted.
Understand, I am playing the devils advocate role more than anything else.
The key point I'm trying to make is that where the data is located matters for whether Apple (or Microsoft or Redhat or whichever company) has the ability or right to read and report on that data.
> at least you are protected from future policy changes if your files in the cloud are encrypted.
If, and only if, that data is never synced back to your phone (which Apples does currently).
The phone would be literally incapable of determining if there were any matches. Please read up on how it was designed to work.
But I think that's pretty thin gruel, since now you're just a feature-flag (or even a bug) away from all content being scanned. More broadly, the entire endeavor is very much at odds w/ the sentiment expressed in their public advertising.
They want to scan on device so they do not have to scan it in the cloud.
Because people are in full FUD mode on this, we're stuck storing photos without full end-to-end encryption because Apple has do the scanning on their side.
> Well, there is another option: Apple could actually respect your privacy, support end-to-end encryption, and not scan your content at all.
This is not an option you, or Apple, think there is to choose.
Peoples FUD and misrepresentation over what was happening and your snark over "yes but their advertising campaign says things stay on the phone" (which is literally true) is IMO likely actually accelerating the privacy degradation. When they are compelled by law to scan this all centrally in one place in the cloud, I guess you will be happier that it's more invasive but just happening "elsewhere"?
If you don't want them to check for unwanted content, don't put your content on their machines.
That’s not true. The system didn’t work the way you are assuming. The device had no knowledge of any matches. The “scanning” was a coöperation of client-side and server-side code, each with an extremely limited knowledge of the data involved.
Apple says this would only be used on iCloud synced photos, but I don't see any technical reason the process could not be performed for any photo.
Am I mistaken?
> The device had no knowledge of any matches.
True, but how is it relevant?
Also the scanning was calculating hashes based on content, not just metadata.
It’s not. The device has no idea if there are any matches. Everybody is assuming how it works without actually reading how it works.
On-device scanning like that would be pointless, though. IIRC, stuff uploaded to their cloud is already accessible to Apple for server-side scanning. The controversial thing was the on-device scans would trigger some kind of upload of un-uploaded stuff to Apple for further investigation.
No, a load of people assumed it would do that, but it’s not possible with the proposed scheme because the device had no knowledge of any matches.
No, the stated behaviour was that it would be uploaded to the cloud at the same time as the encrypted image. Uploaded data was still "encrypted" "at rest" (but possible for them to decrypt, just not done routinely). The whole point of the client-side scanning was that scanning was done locally on the phone and uploaded with it, as an explicit alternative to them regularly decrypting and scanning all user data on their servers. The stated operation was; if the phone fingerprint matched a public CSAM fingerprint list, then it would be automatically decrypted and scanned with a second, private fingerprinting algorithm on the server - only if this second, private fingerprint matched would the data be flagged up for potential violation (and inspected).
The alternative to this was them just regularly scanning all data in their cloud anyway. Apparently people are happier with that option.
Even MEGA, which signals the virtues of privacy/security through the prominence of decryption keys on its UI flows, will still report illegal content to authorities and display a message saying so if content was removed for that reason.
Any publicly traded company that touts perfect privacy cannot deliver what they are claiming, or they'd become the service of choice for every type of disenfranchised person - including child abusers.
Apple has received much more flak than the average corporation over this issue because this fundamental impossibility of perfect privacy clashed with its own privacy signaling in a loud way, and the flurry of debate over the technical merits of the novel, widely shared on-device scanning solution caused much more scrutiny than the boring server-side scanning that has been ubiquitous for decades.
But the fact is that no matter how Apple tries to approach the CSAM problem, it will ultimately have to weed out child abusers from its servers or be publicly and legally lambasted. That is what society has decided is best for the welfare of children, and as a result we will have to live with an imperfect level of privacy as provided by such entities.
That particular issue was the privacy advocacy people run amok.
That means the alternative would be to send the photos to Apple and Apple scans the photo. Either way you send the photo to Apple and meta data gets generated about CSAM. It’s just a matter of where the data gets generated.
I’m also uneasy about it happening on the phone. But honestly, by it being processed on the phone, that means it can be encrypted before it gets to Apple’s servers.
I’m basically working under the assumption that scanning for CSAM is legally required.
It is explicitly not legally required in the US [1]. Providers are required to report "apparent CSAM" that they find on their own, but they are not compelled to search their servers or private devices for its presence.
And this is the case for a very good reason: if it was mandated by US law, then prosecutions would be subject to much stronger 4th amendment review under the "state action doctrine" (i.e., the companies are searching your files without probable cause as compelled representatives of the government.) The current arrangement evades this review under the very thin fig-leaf that US providers are doing the searching on their own.
[1] https://crsreports.congress.gov/product/pdf/LSB/LSB10713
Which would be unfortunate. At that point, you won't be able maintain digital privacy from the govt w/o de-facto becoming a criminal.
CSAM is, I think, simply the initial justification for these systems, since it's widely reviled. But the system itself is not CSAM-specific, and the temptation to expand its scope will likely be irresistible.
If your goal was to become an authoritarian tyrant, you would be very happy to have this in place. :-)
Results about matches aren’t ever on the phone in the first place. Only the server can determine if there are any matches.