Apple clarifies why it abandoned plan to detect CSAM in iCloud photos
wired.com
wired.com
> "It would also inject the potential for a slippery slope of unintended consequences. Scanning for one type of content, for instance, opens the door for bulk surveillance and could create a desire to search other encrypted messaging systems across content types."
Yes, and it was patently obvious from the onset. Why did it take a massive public backlash to actually reason about this? Can we get a promise that future initiatives will be evaluated a bit more critically before crap like this bubbles to the top again? Come on you DO hire bright people, what's your actual problem here?
Whether this was the right response to such concern is something I’m not unsympathetic towards. Certainly I think it’s reasonable to say that Apple was trying to thread a needle in a way which was never going to please everyone, even if it somehow turns out to have been the least-worst outcome.
How could they not see that they would have a giant backlash on their hands? Did they overestimate their ability to get away with the "it's for our children" excuse this badly?"
The ridiculous thing is that Apple's proposal was functionally identical to what other platform vendors (e.g. Google, Microsoft) were already doing. In all cases — including Apple's proposed system — only photos uploaded to cloud storage would be scanned to see if it matched CSAM already known to the government. The only difference with Apple's proposal was initial "fuzzy hash" calculation would be performed on-device prior to upload, instead of on-cloud after upload.
The reason for doing it differently was because it meant (in theory) satisfying both masters — implementing real end-to-end encryption, while not being seen as a CSAM scanning laggard compared to Google, Microsoft, etc.
Other vendors just scan all your shit and nobody cares.
And of the three big platform vendors, two of them already scan private photos for CSAM right now — Google and Microsoft. Yet nobody is outraged because nobody actually cares. There's no logical consistency. Google and Microsoft can implement scanning and there's no outrage. Apple went to great lengths to tell everyone exactly what they were proposing to do before they did it, and all the online people are outraged and calling Apple untrustworthy. Sure, whatever.
Enjoy the exodus.
I reject the claim of "mental gymnastics" as a meme bereft of substance, and specifically in this instance, as an unproven hypothesis. But even if I granted the inference, I don't see how overly elaborate argumentation could have any explanatory power for the trustworthiness (or otherwise) of major platforms.
I'm curious, why do you think I'm advocating trust in any platform? And I'm curious why you think that I have any concern one way or another for an "exodus" from any platform?
Really, I'm just replying to say that while I think I understand the gist of the tone of your reply, I don't actually understand anything you wrote. And my hope is that you can clarify, because — and I say this with all sincerity — I really am genuinely curious.
I imagine it wasn't an environment where one could argue those concerns on fair grounds without it being seen as enabling CSAM and shot down. I also imagine it's enticing to call their competitors' products "pedo-phones".
It worries me that Sarah Gardner seems to be truthful when she hints at that no one at Apple reached out to her after Apple killed the plan in December 2022. She must have also missed articles like the one at The Verge [1].
Gardner’s message to Tim is dated 2023-08-30 03:24:37 (CEST).
Erik Neuenschwander's reply was printed out 2023-08-31 14:41:00 (probably PDT).
So that’s a period of about 44 hours that the print out [0] represents.
It would be helpful to know the time when Apple’s reply was sent to deduce whether Apple had time for additional deliberations before writing the email to Gardner.
Regardless, as reported by The Verge they obviously had deliberations shortly after the public outcry, killed the plan in December 2022, and then presumably failed to inform Gardner about the reasons.
If that’s the case, far smaller companies have checks and balances in place to keep that from happening.
[0]: https://s3.documentcloud.org/documents/23933180/apple-letter...
[1]: https://www.theverge.com/2022/12/9/23500838/apple-csam-plans...
In the old days, and I assume it is still the case, judging from observing Apple for the past 20+ years, every single public response has to go through Apple PR. Even public interview before hand they were given lots of preparation.
It is likely Apple has this prepped for a long time. Or even it was another PR case scenario [1] where the CEO was baited to send this email before a response was given and somehow "leaks" to the press.
[1] Read Submarine PR by Paul Graham.
Do you have a link to that issue? I didn't heard about it (I'm from a third world country)
Post Steve Jobs Apple, especially after Scott Forstall and Katie Cotton left. ( Along with a few other top executives) Tim Cook's Apple were left with people of Harmony. These left aligning, DEI focus people have the same characteristic, their way is for the force of good, hence their way is the only way. Same as early Google in the 00s. ( Privacy is a fundamental human right? Actively securing and promoting Chinese components in their supply chains. )
I am sure CSAM started with good intention. As with most ideals do. But fundamentally they don't work in this complex world.
All roads to hell are paved with good intention.
I hope there are still enough of Steve Jobs' conviction left inside Apple.
Ironic that this opinion is just a reductive and reactionary as the thing that it decries.
When it was made public people shot all sorts of holes both in the technical mechanisms themselves as well as their insufficiency even if they all worked as intended.
We should have some sympathy-- they did have some neat technology. But neat technology isn't enough.
I'm curious why they also withdrew those. For those who don't remember the parental control, which were largely overshadowed by the controversy over the cloud stuff, they were to work like this:
1. If parents had enabled them on their child's device, they would scan incoming messages for sexual material. The scan would be entirely on-device. If such material was found the material would be blocked, the child would be notified that the the message contained material that their parents thought might be harmful, and asked if they wanted to see it anyway.
2. If the child said no, the material would be dropped and that would be the end of it. If the child said yes what happened next depended on the age of the child.
3. If the was at least 13 years old the material would be unblocked and that would be the end of it.
4. If the while was not yet 13 they would be given another warning that their parents think the material might be harmful, and again asked if they want to go ahead and see it. They would be told that if they say "yes" their parents will be notified that they viewed the material.
5. If they say no the material remains blocked and that is the end of it.
6. If they say yes it is unblocked, but their parents are told.
There wasn't a lot of discussion of this, and I only recall seeing one major privacy group object (the EFF, on the grounds that if it reaches step 6 it violates the privacy of the person sending sex stuff to your pre-teen because they probably did not intend for the parents to know).
Not quite the same but see: https://www.forbes.com/sites/thomasbrewster/2023/04/06/sex-t...
Apple’s updated system allows children to ask for help from an adult using the Communication Safety features, which strikes a good balance to me.
uh, by preventing them from seeing sexual content?
“Apple knew my 14 year old was receiving sexual material and did nothing about it?!”
Let's say the parents are abusive, and someone wants to talk with the child about that (via chat, for some reason).
Now, if the algorithms sometimes incorrectly flag private messages that were in fact safe -- could that be mitigated by letting the sender know: "Your message will be scanned and possibly shown to the parents of the recipient" before they hit Send?
(O.t.o.h. that leaks info about the age of the recipient.)
Big fan of keeping different forms of content separate.
It really surprises me a lot of times how little failure analysis is done in software engineering considering it is basically the cornerstone of most physical engineering. It is critical that you design systems to fail in certain ways. That is error messaging... (Also see Blackstone's Ratio for an example in law)
Different forms for different content is a nice answer, but you also have to remember that we're talking a product for the masses. So while I don't think your answer is wrong, it is.
ETA: looks like they directly provide documents from Apple at the bottom of the article
She's using PR to pressure on Apple into implementing the kind of solution her previous company is selling. Won't someone think of the children??
[1] https://www.linkedin.com/in/sarah-gardner-aba90013/ [2] https://www.thorn.org/our-work-to-stop-child-sexual-exploita...
There is only one correct answer though and that is what they have clarified.
I would immediately leave the platform if they progressed with this.
This is not to say they should scan locally, but my understanding of CSAM was that it would only be scanned on its way to the cloud anyways, so users who didn’t use iCloud would’ve never been scanned to begin with.
Their new proposed set of tools seems like a good enough compromise from the original proposal in any case.
To counter the "think of the children" -argument governments use to justify surveillance, Apple tried scanning stuff on-device but the internet got a collective hissy-fit of intentionally misunderstanding the feature and it was quickly scrapped.
They basically did. If you turn on Advanced Data Protection, you get all of the encryption benefits, sans scanning. The interesting thing is that if you turn on ADP though, binary file hashes are unencrypted on iCloud, which would theoretically allow someone to ask for those hashes in a legal request. But it's obviously not as useful for CSAM detection, as, say, PhotoDNA hashes. See: https://support.apple.com/en-us/HT202303
For anyone else wondering, to enable it just go to iOS Settings -> iCloud and you'll see "Advanced Data Protection." Toggle it to enabled to create a recovery key, which you'll then be prompted to input correctly after saving it somewhere safe, and then return to the iCloud Settings page, toggle it one more time and enter your recovery key again to confirm.
> scanning stuff on-device
What do you think they were going to do once the scanning turned up a hit? Access the photos? Well that negates the first statement.
In the whitepaper, the cryptography required that Apple have multiple different photodna (or whatever the name was for the on-device one) matches before they could unwrap the user's message containing these suspected CSAM photos and to then send them to NCMEC.
So the resolution most likely would've been the same, but the detail blurred so that the poor human agent wouldn't have to see actual CSAM, just enough to make a call whether it is or isn't a likely match.
https://www.apple.com/child-safety/pdf/CSAM_Detection_Techni...
Once you reached a certain threshold (the number was not given) it would trigger an alert in a system at Apple.
Each report contained a bit of data that wasn’t enough to identify someone. Once enough “points” from one account accumulated they’d have enough to identify who you were, which files matched, and presumably the full decryption key.
I believe the plan was the suspect files would be decrypted and compared against the real CSAM signatures. If a close match was found it would be sent to NCMEC for confirmation and law enforcement actions.
The threshold was to prevent false positives from the perceptual hashes, like the Google AI scanning incident. Reportedly nobody has one or two pictures. People with CSAM tend to have a lot, so they’d show up “bright red”. They probably didn’t want to reveal the number so people wouldn’t try to keep only that many pictures on their phone to avoid detection.
This was all outlined in the first two pages of the white paper, and explained in more detail further down.
https://www.apple.com/child-safety/pdf/CSAM_Detection_Techni...
how was it misunderstood? your device would scan your photos and notify apple or whoever if something evil was found. wasn't that what they were trying to do?
Whereas right now all the other cloud providers just send the photo as-is and scan it on the cloud servers.
With Apple's approach, the cloud servers don't get to look at every single one of your photos like cloud vendors do today, scanning happens within the privacy of your own phone, and only known-kiddy-porn signatures are flagged.
Apple came up with a way to make things way more private, but the concept of your own device working "against" you if you happen to be a pedophile was too much of a leap.
And it wouldn't have notified Apple of "something evil", just specifically known and human-verified actual real child abuse photos. And not even that, it would have needed multiple matches of those very real and verified abuse photos before it flagged them so that a real human could see a "visual derivative" of the photos.
Only if those multiple matches of derivatives were deemed as actual, very real, child pornography the authorities would've been called.
But nope. Now they just scan ALL your data in the cloud when the authorities demand it. And that's somehow better according to the internet in a way I still can't understand.
I don’t see the problem with this status quo. There is a clear demarcation between my device and their server. Each serving the interests of their owner. If I have a problem with their policy, I can choose not to entrust my data to them. And luckily, the data storage space has heaps of competitive options.
The generic space does, yes. But if you want native integration with iOS, your only choice is iCloud. It would certainly be nice if this was an open protocol where you could choose your own storage backend. But I think the chances of that ever happening are pretty much zero.
2. The division you envisage (“The software running on the phone should be representing the owner of the phone, period.”) is wishful thinking. Do you think the JavaScript in your browser does only things in your interest?
Web javascript is one of the places the battle is being fought. Users are being pushed into running javascript (and HTML) that acts directly against our own interests (eg ads, surveillance, etc). Many of the capabilities exploited by the hostile code should be considered browser security vulnerabilities, but the dynamic is not helped by one of the main surveillance companies also making one of the main browsers.
But regardless of the regime the authoritarians are trying to push, the computer-represents-user model is what we should aspire to - the alternative is computational disenfranchisement.
In this specific case it is not wishful thinking.
The feature got scrapped. Users and people who support privacy won.
It’s objectively better than what google does but I’m glad we somehow ended up with no scanning at all.
Apple doesn’t want to expand their power which is why they don’t scan locally. They weren’t doing it before and they don’t want to offer it now.
The policy is meant to ensure Apple's servers are not storing and distributing CSAM, not that Apple wants to become a police investigative force.
No outright statement confirming or denying this has ever made to my knowledge, but the implication, based both on Apple's statements and the statement of stakeholders, is that this isn't currently the case.
This might come as a surprise to some, because many companies scan for CSAM, but that's done voluntarily because the government can't force companies to scan for CSAM.
This is because based on case law, companies forced to scan for CSAM would be considered deputized and thus subsequently it would be a breach of the 4th amendments safeguards against "unreasonable search and seizure".
The best the government can do is to force companies to report "apparent violations" of CSAM laws, this seems like a distinction without a difference, but the difference is between required to actively search for it (and thus becoming deputized) v. reporting when you come across it.
Even then, the reporting requirement is constructed in such a way as to avoid any possible 4th amendment issues. Companies aren't required to report it to the DOJ, but rather to the NCMEC.
The NCMEC is a semi-government organization, autonomous from the DOJ, albeit almost wholly funded by the DOJ, and they are the ones that subsequently report CSAM violations to the DOJ.
The NCMEC is also the organization that maintains the CSAM database and provides the hashes that companies, who voluntarily scan for CSAM, use.
This construction has proven to be pretty solid against 4th amendment concerns, as courts have historically found that this separation between companies and the DOJ and the fact that only confirmed CSAM making its way to the DOJ after review by the NCMEC, creates enough of a distance between the DOJ and the act of searching through a person's data, that there aren't any 4th amendment concerns.
The Congressional Research Service did a write up on this last year for the ones that are interested in it[0].
Circling back to Apple, as it stands there's nothing indicating that they already scan for CSAM server-side and most comments both by Apple and child safety organizations seem to imply that this in fact is currently not happening.
Apple's main concerns however, as stated in the letter by Apple, echo the same concerns by security experts back when this was being discussed. Namely that it creates a target for malicious actors, that it is technically not feasible to create a system that can never be reconfigured to scan for non-CSAM material and that governments could pressure/regulate it to reconfigure it for other materials as well (and place a gag order on them, prohibiting them to inform users of this).
At the time, some of these arguments were brushed off as slippery slope FUD, and then the UK started considering something that would defy the limits of even the most cynical security researcher's nightmare, namely a de facto ban on security updates if it just so happens that the UK's intelligence services and law enforcement services are currently exploiting the security flaw that the update aims to patch.
Which is what Apple references in their response.
> Nothing in this section shall be construed to require a provider to—
> (1) monitor any user, subscriber, or customer of that provider;
> (2) monitor the content of any communication of any person described in paragraph (1); or
> (3) affirmatively search, screen, or scan for facts or circumstances described in sections (a) and (b).
The core of 18 U.S. Code § 2258A - Reporting requirements of providers is available at https://www.law.cornell.edu/uscode/text/18/2258A.
Great addition to provide more context.
>(e) Failure To Report.—A provider that knowingly and willfully fails to make a report required under subsection (a)(1) shall be fined— (1) in the case of an initial knowing and willful failure to make a report, not more than $150,000; and (2) in the case of any second or subsequent knowing and willful failure to make a report, not more than $300,000.
I find these clauses at odds with one another in that the Failure to Report clause created a tangible duty upon the provider, which, were I a judge, would satisfy me that rhe provider was, in fact, deputized.
Does nobody actually read the legislation that is passed and realize that oops, I just passed am unconstitutional law.
That they include the construed... clause just solidifies for me that the legislators in question were trying to pull a fast one.
Put simply, if they have knowledge of it they have a duty to report, but they can’t be compelled to try and find out.
In theory this means that if they happen to stumble upon it or are being alerted to it by a third party (e.g. user report) then they have to report it, in practice many voluntarily monitor it, maybe because they want to avoid having to litigate that they didn’t have knowledge of it or maybe because it’s good PR or maybe because they care for the case.
I think in most cases it’s all of the above in one degree or another.
Words mean things, and we've gotten damned loose with it these days in my opinion when the want strikes. "Voluntary" anything with a $150000 fine for not doing it is no longer voluntary. It's now your job. If it's your job, and the State punishes you for not doing it, you are a deputy of the State. I do not care how many layers of legal fiction and indirection are between you and the State.
If you can't not comply without jeopardy, it ain't voluntary.
Absolutely not. That section requires a report under the circumstances where a provider has obtained “actual knowledge of facts and circumstances” of an “apparent violation” of various code sections (child porn among others). It doesn’t place on the provider the burden of seeking out that knowledge. In other words, it covers the cases where, for example, a provider receives a report that they are hosting a child porn video and are pointed to the link to it. Providers can’t jam their fingers in their ears and shout LALALA when they’re told they’re hosting (or whatever) CSAM and given the evidence to support it. They don’t have to do anything at all to proactively find it and report it, however.
Think of it like this. I, as a high school, teacher, am a mandated reporter of child abuse. It’s literally a crime (a misdemeanor) for me not to report suspected child abuse. But I don’t have to go out and suss out whether any of my students are being abused. That doesn’t make me a state actor for 4th Amendment purposes (although I am otherwise, because I am a public school teacher, but that’s a different issue).
A spade, is a spade by any other name. If the state will come down on you for not doing something (message generation), you are a deputy of the State. Period.
Since it all went down they added the advanced security option that encrypts photos, messages, and even more.
But that option is opt-in since if you mess it up they can’t help you recover.
That said, I could be wrong about them not scanning currently, I simply don’t have anything authoritative saying either way.
Only statements that imply that they currently don’t, nothing more.
I speculated (and now we know) when this new scanning announced, that it was in preparation for full E2EE. Apple came up with a privacy preserving method of trying to keep CSAM off their servers while also giving E2EE.
The larger community arguments swayed Apple from going forward with their new detection method, but did not stop them from moving forward with E2EE. At the end of the day they put the responsibility back on governments to pass laws around encryption - where they should be, though we may not like the outcome.
At the time I also thought it was obvious it was in preparation for e2ee (despite loud people on HN who disagreed).
I do wonder if they had intended to have it be default on though, maybe not since probably better for most users to have a recovery option.
By definition, encryption (with unique user keys) means you can't infer nor check what the content of the message is. Not without client cooperation, which is what this feature would have been.
> “Convergent encryption solves this problem in a very clever way:
“The way to make sure that every unique user with the same file ends up with an encrypted version of that file that is also identical is to ensure they use the same key. However, you can’t share keys between users, because that defeats the entire point; you need a common reference point between users that is unknown to anyone but those users.
“The answer is to use the file itself: the system creates a hash of the file’s content, and that hash (a long string of characters derived from a known algorithm) is the key that is used to encrypt said file.
“If every iCloud user uses this technique — and given that Apple implements the system, they do — then every iCloud user with the same file will produce the same encrypted file, given that they are using the same key (which is derived from the file itself); that means that Apple only needs to store one version of that file even as it makes said file available to everyone who “uploaded” it (in truth, because iCloud integration goes down to the device, the file is probably never actually uploaded at all — Apple just includes a reference to the file that already exists on its servers, thus saving a huge amount of money on both storage costs and bandwidth).
“There is one huge flaw in convergent encryption, however, called “confirmation of file”: if you know the original file you by definition can identify the encrypted version of that file (because the key is derived from the file itself). When it comes to CSAM, though, this flaw is a feature: because Apple uses convergent encryption for its end-to-end encryption it can by definition do server-side scanning of files and exploit the “confirmation of file” flaw to confirm if CSAM exists, and, by extension, who “uploaded” it. Apple’s extremely low rates of CSAM reporting suggest that the company is not currently pursuing this approach, but it is the most obvious way to scan for CSAM given it has abandoned its on-device plan.”
https://stratechery.com/2022/apple-icloud-encryption-csam-sc...
So every file has a unique key? So thousands or 10,000s of keys would need need to be in the keychain, mapped to the file name.
And if one person's keys are leaked, they can be used prove that other people had the same file
No, this doesn't sound well thought out
The file encryption part was based on using a hash of the file as the key.
It's always nice to later find out that one's quick amateur idea turns out to be an independent rediscovery of something legit. Now that I've learned it is called "convergent encryption" Googling tells me it it goes back to 1995 and a Stac patent.
For regular people taking photos the government won’t have their plaintext.
For popular media people are uploading the same copy of it saves a lot of bandwidth.
I thought the same.
> despite loud people on HN who disagreed
Yeah, loud people be like that, but this is really Apple’s communication fault. They could have started with that “hey we want to provide e2e encrypted storage, the price of it will be that we need to scan what you upload for csam”.
Companies don't want to employ people. People are annoying. They make annoying demands like wanting time off and having enough money to not be homeless or starving. AI should be a tool that enhances the productivity of a worker rather than replacing them.
Fully automated "safety" systems always get weaponized. This is really apparent on Tiktok where reporting users you don't like is clearly brigaded becasue a certain number of reports in a given period triggers automatic takedowns and bans regardless of assurances there is human review (there isn't). It's so incredibly obvious when you see a duet with a threatening video gets taken down while the original video doesn't (with reports showing "No violation").
Additionally, companies like to just ban your account with absolutely no explanation, accountability, right to review or right to appeal. Again, all those things would require employing people.
False positives can be incredibly damaging. Not only could this result in your account being banned (possibly with the loss of all your photos on something like iCloud/iPhotos) but it may get you in trouble with law enforcement.
Don't believe me? Hertz was falsely reported their cars being stolen [1], which created massive problems for those affected. In a better world, Hertz executives would be in prison for making false police reports (which, for you and me, is a crime) but that will never happen to executives.
It still requires human review to identify offending content. Mass shootings have been live streamed. No automatic system is going to be able to accurately differentiate between this and, say, a movie scene. I guarantee you any automated system will have similar problems differentiating between actual CSAM and, say, a child in the bath or at the beach.
These companies don't want to solve these problems. They simply want legal and PR cover for appearing to solve them, consequences be damned.
[1]: https://www.npr.org/2022/12/06/1140998674/hertz-false-accusa...
The whole uproar about this system was made by people who didn't know the most basic things about it.
The tl;dr is that despite this man ultimately having his name cleared by the police after having his entire Google account history (not just cloud) searched as well his logs from a warrant served to ISP, Google closed his account when the alleged CSAM was detected and never reinstated it. He lost his emails, cloud pictures, phone number (which losing access to prevented the police from contacting him via phone), and more all while going through a gross, massive invasion of his privacy because he was trying to do right for his child during a time when face-to-face doctor appointments were difficult to come by.
This should be a particularly salient reminder to people to self-host at the very least the domain for their primary and professional e-mail.
[0] https://www.nytimes.com/2022/08/21/technology/google-surveil...
The google one actually does try to detect new ones and there are reported instances of Google sending the police on normal parents for photos they took for the doctor.
Pick your favourite other example of when Apple and Google have faced roughly the same problem as each other, and hold up their respective solutions next to those in the example of CSAM scanning above. I bet they'll look similar.
And that's ignoring the fact that his photo wouldn't have matched anyway because it isn't in any CSAM database.
And there's another dynamic where telling your customers you're going to scan their content for child porn is the same as saying you suspect your customers of having child porn. And your average non-criminal customer's reaction to that is not positive for multiple reasons.
> (2) Civil liability
> No provider or user of an interactive computer service shall be held liable on account of—
> (A) any action voluntarily taken in good faith to restrict access to or availability of material that the provider or user considers to be obscene, lewd, lascivious, filthy, excessively violent, harassing, or otherwise objectionable, whether or not such material is constitutionally protected
Judges and juries.
And I can assure you that every single judge in the USA, and almost every single member of a jury would decide that CSAM is obscene.
Thats how the law works. We have tons of laws that use general words like this, and trying to be "clever" usually just results in a lost court case or prison time for the person who thinks they found a loophole.
Also you didn't define 'otherwise objectionable'.
For example what I think is excessive violence may be considered normal by a jury in the US...
Quite a lot of things may or may not fall under that definition. Thats how the law works.
> And why would I have to accept an US jury's opinion?
Well, because they and judges are the ones empowered by the government monopoly on violence to judge the law and then have it be enforced, thats why.
Ignore the law at your own peril. But anyway, even if you did ignore the law, this doesn't have anything to do with you.
This is about companies immunity from prosecution. So, even if you disagree with the law, those companies are still immune under section 230, for good faith efforts to remove obscene content.
> Also you didn't define 'otherwise objectionable'.
It would be defined as whatever judges and juries define it as. I don't define it. Instead it is defined by those people.
> For example what I think
If you are not a judge or currently on a jury, then what you think is irrelevant.
The law is not computer code. Instead, it is interpreted by humans. And that is the case for basically all of law.
... by the US government on US territory, i think. They have no business defining "otherwise objectionable" elsewhere.
The principles that I have describe also apply to other countries as well.
So yes it is absolutely the case that in other countries there are judges and juries that apply the law via the process of judges and juries.
But once again, no matter what some other country thinks, on this specific topic, it is about a company getting immunity from USA enforcement. So all other countries do not matter on this topic, because this is about US law.
Both of these arguments are absolutely, unambiguously, correct.
The other side of the coin is that criminals are using E2EE communication systems to share sexual abuse material in ways and at rates which they were not previously able to. This is, I argue, a bad thing. Is is bad for the individuals who are re-victimised on every share. It is also bad for the fabric of society at large, in the sense that if we don't clearly take a stand against abhorrent behaviour then we are in some sense condoning it.
Does the tech industry have any alternate solutions that could functionally mitigate this abuse? Does the industry feel that it has any responsibility at all to do so? Or do we all just shout "yay, individual freedom wins again!" and forget about the actual problem that this (misguided) initiative was originally aimed at?
Cops need to investigate the same way they always have, look for clues, go undercover, infiltrate, find where this stuff is actually being made, etc.
Scanning everyone's phones would make their jobs significantly easier, no doubt, but it simply isn't worth the cost to us as a society and there is simply no good counter-argument to that.
"Apple" wasn't scanning your phone, neither was there a "backdoor".
If you would've had iCloud upload enabled (you'd be uploading all your photos to Apple's server, a place where they could scan ALL of your media anyway), the phone would've downloaded a set of hashes of KNOWN and HUMAN VERIFIED photos and videos of sexual abuse material. [1]
After THREE matches of known and checked CSAM, a check done 100% on-device with zero data moving anywhere, a "reduced-quality copy" would've been sent to a human for verification. If it was someone sending you hashbombs of intentional false matches or an innocuous pic that matched because some mathematical anomaly, the actual human would notice this instantly and no action would've been taken.
...but I still think I was the only HNer who actually read Apple's spec and just didn't go with Twitter hot-takes, so I'm fighting windmills over here.
Yes, there is always the risk that an authoritarian government could force Apple to insert checks for stuff other than CSAM to the downloaded database. But the exact same risk exists when you upload stuff to the cloud anyway and on an even bigger scale. (see point above about local checks not being enabled unless iCloud sync is enabled)
[1] It wasn't an SHA-1 hash where changing a single bit in the source would make the hash invalid, the people doing that were actually competent.
Yes, you're right. But I've seen calls for phones to scan all content before being uploaded or encrypted, and it often feels, at least in some countries, that could still plausible happen. I suppose that's what I had in mind when I wrote my comment.
> But the exact same risk exists when you upload stuff to the cloud anyway and on an even bigger scale.
There's a difference with them actively doing it and announcing they are doing it, vs the possibility they are doing it silently without consent.
It's a cost of doing business pretty much, you need to give the authorites access to customers data or they can go "but think of the children, there might be child abuse material in there!" and that's really damn hard to argue against.
Thus: checking stuff on-device and keeping the cloud locked so that not even the hoster can access it nor can they create a backdoor.
At least some cloud providers require a warrant before doing so.
But the point is that if the data is fully encrypted, no warrant will help against pure mathematics. A cloud provider cannot give something they have no access to.
Right, but your post kind of made it seem like they didn't. And with the CSASM stuff, they didn't need anything to match hashes. So what happens when that gets expanded to other types of content that they still don't need a warrant for? Like torrent files?
> But the point is that if the data is fully encrypted, no warrant will help against pure mathematics. A cloud provider cannot give something they have no access to.
Agreed, but most don't use encryption, and should still have privacy rights for their data.
The point of the system would've been that EVERYONE'S content would've been encrypted in the cloud without them needing to do anything.
also: If it was someone sending you hashbombs of intentional false matches or an innocuous pic that matched because some mathematical anomaly, the actual human would notice this instantly and no action would've been taken. - if someone is doing this, imagine the scale- thousands of pics that should be human-evaluated, scaled to thousands of people, it'll be just plain ignored, meaning system loses it's purpose. also, you say that it'll be enabled only if icloud bck is enabled, but it's not guaranteed, this assumption can later change... and it doesn't make sense, for me your 2 statements contradict themselves:
- if apple can scan your photos in icloud AND for this feature to be enabled, you must enable icloud, why they should send hashes to you? they can scan the photos anyway in icloud, since all your photos are backed up there. Unless... they can't scan photos in icloud since these are encrypted, meaning scanning can be done only locally before photos are sent, meaning icloud enabling is not mandatory and it could work without it.
Either way the csam scanning is imo pointless, on one hand bc of privacy reasons(and we've seen that if state is able to use a backdoor, it'll use it when needed) and on the other hand, because of generative algorithms: photos can be manipulated to trigger csam even if human eye can see another thing (aka hashbomb) OR a sick/ill intentioned person can generate a legit csam like photo just by using target ppl's face(or description of their face), in this case I don't even know if they are breaking the law or not, since the image is totally generated but is looking totally illegal
"Looking illegal" isn't enough to trigger a CSAM check in this case. It's perfectly normal to take pictures of your own kids without clothes in most of Europe for example. Nothing illegal.
That's why the checks would've been explicitly done on known and confirmed CSAM images. It wasn't some kind of check_if_penis() -algorigthm, or one of the shitty ones that trigger if there's too much (white) skin colour in an image.
Outsourcing work for this(afaik) isn't possible since it's private data, not public and only specific organisations can have full access to potential triggers
But in the end it also doesn't matter because there are other problems too, like how to make the final list easily checkable so that we are sure governments/ companies do not alter the list to target specific ppl/groups for their own interest. Or how algorithm isn't modified under the hood to check not just images but also text/files
Nobody read the bit about an actual human verifying results before any law enforcement would be called in.
And outsourcing checking is a huge industry even today[0]. How do you think the huge social media companies keep CSAM, gore etc out of their systems? They're not using Pied Piper's hotdog or not algorithm, that's for sure.
[0] https://www.theverge.com/2019/2/25/18229714/cognizant-facebo...
A snippet from the article:
> The video depicts a man being murdered. Someone is stabbing him, dozens of times, while he screams and begs for his life. Chloe’s job is to tell the room whether this post should be removed.
Nobody read the bit about an actual human verifying results before any law enforcement would be called in. - again, if system can be gamed, human check is useless. Imagine someone will generate 100k false positives, multiply this by 1k ppl and imagine among them there is a real bad person with 10 real csam images and another person with fully generated csam images (like generated porn is a thing now so generating csam like images is possible). How do you think gov will human evaluate 100kk pictures that are triggering the system? Because if they can't, system is useless for this usecase but is still useful for potential gov oppression or company ad targeting
You can generate a billion false positives and it still won't do anything. You need to get them to people's iCloud photo libraries first. Each library needs to have multiple false positive images before triggering a human check. They intentionally didn't tell the exact number needed, but it's not two.
If you have a way to get enough people to get fake CSAM material on their phones to overwhelm the human checkers, why would you waste in on something stupid like that? Just by forcibly inserting advertisements to peoples Photo Libraries would make you a billionaire. Not an ethical one, but still rich.
Oh, and just for reference. FB gets 350 million photos uploaded every day and they keep it moderated just fine. The amount of people you'd need colluding with you to overwhelm the system Apple had designed would be staggering.
And then you've achieved what? Make it possible to share child pornography because you broke the system? Yay, victory?
The local scanning would've been enabled only if you would've uploaded the photos to iCloud anyway.
To complete your crappy analogy: You had already agreed to send the information of everything contained in your apartment to a company, and now you're getting into a hissy-fit when they are sending a robot visit your house with a list of known child pornography photos and checking if you have any?
There weren't any "hard decisions" it wasn't a "this has a naked person in it" -scanner. It wasn't an "abuse checked" it was specifically created to find known CP.
It checked your photos against a NeuralHash of KNOWN child pornography. So the only chance your sauna photos would've been flagged if they were included in a set of CP photos distributed widely enough so that they're added to the CSAM database.
And nobody claimed the system would magically cure the world of grooming, where did you get that idea from? Although the current filters in iMessage might help if the abuser is stupid enough to use that for grooming.
Yes, i know they scan all photos already. But to scan MY photos and hold me accountable, that is something i am not okay with. Yes, its only icloud, for now, but i dont see the point of using some 3rd party cloud, i used to, but in apple device it would be suboptimal for my daily use etc etc.
My point was of the original case of protecting children. That is what the advocate groups want. That is what I want. But I just draw the line there what i feel is my personal space, that you don't scan my photos or files, and that is not the answer. Since there are these few behemoth FAANG companies ruling the ecosystems and impacting our lives, the battle of privacy needs to be fought right there. So, for me that would be the line as I see the progression of that path is much worse. That is where I raise my hand and say, I'm out, what next? No smartphone? Maybe. :) Peace.
They can find those materials the same way abusers work their way into communities to have access to them in the first place.
The increased scale only means they need more people working on it.
That aside, we could also, at least in the US, stop taking such a ridiculous stance against drugs and instead prioritize finding makers of CSASM materials.
The people are there, they are just not being utilized effectively.
It hasn't particularly changed production, which is where the actual abuse happens. There are still actual people abusing and filming actual children, and those can be found by the police by the same old-fashioned methods they've always had available. (Plus many new ones that don't violate everyone's civil liberties or destroy the security of every networked device.)
I'm all for privacy, but those who put it above all else are already likely not using Apple devices because of the lack of control. I feel like for Apple's target market the implementation was reasonable.
I think Apple backed down on it because of the vocal minority of privacy zealots (for want of a better term) decided it wasn't the right set of trade-offs for them. Given Apple's aim to be a leader in privacy they had to appease this group. I think that community provides a lot of value and oversight, and I broadly agree with their views, but in this case it feels like we lost a big win on the fight against CSAM in order to gain minor, theoretical benefits for user privacy.
Idk, maybe i'm overreacting, but I've seen what a repressive gov can do, and with such an instrument it's frightening what surveillance vectors can be opened
I don't belive tech has an over weighted responsibility to solve society's problems, and in fact it's generally better if we don't try and pretend more tech is the answer.
Advocating for more money and more prioritization for this area of law enforcement is still the way to go if it's a priority area. Policing seems to be drifting towards "mall cop" work, giving easy fines, enabled by lazy electronic surveillance casting a wide net. Let's put resources towards actual detective work.
Apple's proposed solution would have theoretically only reported cases that were much more than likely to be already known instances of CSAM (i.e. not pictures of your kids), and if nothing else is reported, can we say that they were really surveilled? In some very strict sense, yes, but in terms of outcomes, no.
Also, you can not guarantee that apple/google will use only known instances of csam, what if, govt orders them/google to scan for other type of content under the hood, like documents or god knows what else bc govt want's to screw that person (for the sake of example let's suppose the targeted person is some journalist that discovered shady stuff and govt wants to put em in prison), bc you know, you don't have access to either algorithms and csam scan list that they are using, system could be abused and usually could means 'sometime' it will
I agree that the basic idea of scanning on device for CSAM has a lot of issues and should not be implemented. What I think was missing from the discourse was an actual look at what Apple were suggesting, in terms of technical specifics, and why that would be well designed to not suffer from these problems.
https://www.texastribune.org/2022/07/26/texas-foster-care-ch...
So how about we implement mass-surveillance by giving victims a good reason to report crimes? Starting with not heavily punishing victims that do come forward. Make the foster care system actually able to raise kids reasonably.
Because, frankly, if we don't do it this way, what's the point? Why would we do anything about abuse if we don't fix this FIRST? Are we really going to catch sexual abuse, then put the kids into a state-administered system ... where they're sexually, and physically, and mentally, and financially abused?
WHY would you do that? Obviously that doesn't protect children, it only hides abuse, it protects perpetrators in trade for allowing society to pretend the problem is smaller than it is.
In terms of outcomes, almost nobody is actually surveilled, as the overall effect is the same as no data having been collected on them in the first place.
That said, I am personally more comfortable with my country's intelligence agencies hoovering up all my online activity than I am with the likes of Apple. The former is much more accountable than the latter.
What if you were a candidate for political office, pushing opinions that angered large swaths of the Intelligence Comminity?
The "minuscule fraction" of content is not surfaced by some random roll of the dice - it's the definitionally most interesting content, in the sense that some human went specifically looking for it in the heap of content caught in the dragnet. And it only needs to be interesting to at least one person with the clearance to search for it.
Maybe that means it's a video of a child being abused, and some morally upstanding federal officer is searching for it because anyone possessing it is ethically and legally culpable for the abuse of that child... Or maybe it's a PDF containing evidence of FBI kidnapping and torturing innocent civilians, and some morally corrupt federal officer is searching for it because anyone possessing it is a liability who needs to be silenced... Or maybe it's a JSON file containing the GPS locations of an individual for the past year, and some emotionally scorned federal contractor is searching for it because that individual is their ex-spouse who's moved onto a new partner.
Are you really prepared to put your faith in the trustworthiness and moral clarity of the population of 100k+ people with federal security clearances?
The scenarios you invented sound very far-fetched to me, if these did happen I very much doubt the perpetrator would be able to get away with it.
https://www.eff.org/deeplinks/2023/04/internal-documents-sho...
> At least a dozen U.S. National Security Agency employees have been caught using secret government surveillance tools to spy on the emails or phone calls of their current or former spouses and lovers in the past decade, according to the intelligence agency’s internal watchdog.
https://www.reuters.com/article/us-usa-surveillance-watchdog...
We deceive ourselves honestly by pretending like we have not created new realities which are problematic at scale. We have. They are plentiful. And if people aren’t willing that we walk back tech to reduce the problems and people aren’t willing to accept technical solutions which are invasive then what are we to do? Are we just to accept a new world with all these problems stemming from unintended consequences of tech?
Enabling people to talk to each other without all their communication being wiretapped and archived forever is not one of those, I would say.
Uber and AirBnB are just using very-widely-available technology—that some taxi services and hotels are also using!—and claiming that they're completely different when the main difference is that they're just ignoring the laws and regulations around their industries.
Cryptocurrencies are using a tech innovation as a front for what's 99.9999% a financial "innovation"...which is really just a Ponzi scheme and/or related scams in sheep's clothing.
LLMs are genuinely a tech innovation, but the primary problem they bring to the fore is really a conversation we've needed to have for a while about copying in the digital age. The signs have been there for some time that such a shift was coming; the only question was exactly when.
In none of these cases is technology actually doing anything "messed up". Companies that denote themselves as being "in the tech industry" do bad things all the time, but blaming the technology for the corporate (and otherwise human) malfeasance is very unhelpful. In particular, trying to limit technological progress, or ban widely useful technological innovations because a small minority of people use them for ill, is horrifically counterproductive.
Enforce the laws we have better, be more willing to turn the screws on people even if they have lots of money, and where necessary put new regulations on human and corporate behavior in place (eg, requiring informed consent to have works you created included in the training set of an LLM or similar model).
You haven't explained the problem 'tech' has created, I'm confused as to what your point is?
CSAM isn't a problem caused by 'tech' unless you're going back to the invention of the camera, and I think that toothpaste is well out of the tube.
Additionally, and this is where a whole of arguments about this go wrong, the important part: the actual literal abuse, is human to human. There is no technology involved whatsoever.
Technological involvement may be an escalation of offense, but it's vanishingly secondary.
If illegal data (CP) is being transferred on the net, wiretapping that traffic and bringing hits to the attention of a human seems like a proportional response.
(Yes, I know, it's not going to be 100% effective, encryption etc, but neither is actual detective work.)
For this I need proof.
But it’s just begging the question to say there’s immense benefit to them searching US citizens’ communications without a reason.
That’s the whole question.
Show me why we should change the constitution which guarantees us freedom from this sort of government oppression.
Some of this will include data of British citizens, but the thing is, we have a significant home-grown terrorism problem and serious organised criminal gang activity, happening within the country. If intelligence analysts need to look at, for example, which phone number contacted which other phone number on a specific date in the recent past, there's no other way to do this other than bulk collect all phone call metadata from the various telecom operators, and store it ready for searching.
The vast majority of that data will never be seen by human eyes, only indexed and searched by automated systems. All my phone calls and internet activity will be in there somewhere, I'm sure, but I don't consider that in itself to be government oppression. Only if it's used for oppressive purposes, would it become oppressive.
Congrats, you've got a religion.
If you want to get it amended, then by all means, make a case for why it should be amended.
In the meantime you wanted to know why mass surveillance isn’t an option. The answer “because it’s against the law” is a simple, good answer.
If you want to know why we decided as a nation to make that such a fundamental law that it is in our constitution, you could do worse than reading about what prompted the writing of the Bill of Rights.
I agree with a lot of the original reasoning.
While often true, at all times there have also been morally wrong laws, so it would not be unreasonable to counter that being written into law on itself means nothing. So you should always be prepared to pull out and defend the reasoning behind a law, which you also hinted at in your following sentences.
I think we’re well beyond that point now. Whether or not encryption is allowed or not and however private you believe your virtual life to be, in the physical world surveillance is the norm. Your physical location, biometric information, and relationships can and will be monitored and recorded.
> Does the tech industry have any alternate solutions that could functionally mitigate this abuse? Does the industry feel that it has any responsibility at all to do so? Or do we all just shout "yay, individual freedom wins again!" and forget about the actual problem that this (misguided) initiative was originally aimed at?
The issue at large here is not the tech industry, but law enforcement agencies and the correctional system. Law enforcement has proven time and time again themselves that the most effective way to apprehend large criminal networks in this area is by undercover investigation.
So no, I don't think it is the tech industries role to play the extended arm of some ill conceived surveillance state. Because Apple is right: This is a slippery slope and anyone that doesn't think malicious political actors will use this as a foot in the door to argue for more invasive surveillance measures, using this exact pre-filtering technology are just naive idiots, in my oppinion.
Much like other 'tough on crime' measures (of which destroying E2EE is one) the real problems need to be solved not at the point of consumption (drugs, guns, gangs, cartels) but at the root causes. Getting rid of E2EE just opens the avenue for the abuse of us by the government but in no way guarantees we'll meaningfully make children safer.
And no, we are not 'condoning' it when we declare E2EE an overall good thing. Real life is about tradeoffs not absolutes, and the tradeoff here is protection from the government for potentially billions vs. maybe arresting a few thousand more real criminals. This is a standard utilitarian tradeoff that 'condones' nothing.
Governments are not good at this type of thing. It requires careful analysis, planning and actual decisions.
But slap on a regulation and require private companies to do the hard work for you - now we are talking!
Yes. Since becoming an abuser is a process and not a moment, part of the solution must be making access to CSAM much harder.
> And no, we are not 'condoning' it when we declare E2EE an overall good thing.
Agreed. I'm sorry if I worded things in a way that caused you to see an implication which was not intended. To be clear: E2EE is a good thing. Championing E2EE is not equivalent to condoning CSAM.
What I did say is that in failing to try and provide any meaningful solutions to this unintended consequence of E2EE, the industry is effectively condoning the problem.
> This is a standard utilitarian tradeoff
If that's the best we can do, I'm very disappointed. That position says that to achieve privacy, I must tolerate CSAM. I want both privacy and for us not to tolerate CSAM. I don't know what the solution is, but that is what I wish the industry were aiming for. At the moment, the industry seems to be aiming for nothing but a shrug of the shoulders.
Not true, you can have privacy and at the same time not tolerate child pornography, those are two perfectly compatible positions and arguably the current state. What you can not have - by definition - is privacy on the one hand and on the other hand no privacy in order to look for child pornography. You can still fight child pornography in any other way, but when it comes to privacy, you have to make a choice - give people their privacy or look through their stuff for illegal content, you can not have both. If you have enough evidence, a judge might even grant law enforcement the permission for privacy violating measures, it should just not be the default position that your privacy gets violated.
Failure to recognize this perpetuates the fundamental inconsistency.
[1] https://en.wikipedia.org/wiki/Relationship_between_child_por...
I want to have more money than Elon Musk..... sometimes life is not fair and we can not always get what we want...
Any "backdoor" or "frontdoor" in encryption is a total failure of encryption. That is a immutable truth, more fixed in reality than the speed of light is in physics.
If you can't even propose a hypothetical science fiction-esque way of achieving this, much less one that might actually be implementable now or in the near future (like, 5-10 years), you shouldn't get to have that opinion.
In my opinion CSAM is a symptom, not a cause.
It's difficult to "stumble across" that kind of material unless you're already actively looking for it, which means some amount of "damage" is already done.
I also highly doubt that someone with no proclivities in that direction would 'turn' as a result of stumbling across CSAM. I'd guess they'd go the other way and be increasingly horrified by it.
It's entirely likely that borderline cases (15-17 years old) are seen by millions of people without them realizing it. Pornhub is a popular "mainstream" porn website that has issues with CSAM and routinely removes it when found. It's entirely possible for "normal" consumers of pornographic material to stumble into CSAM in places like that unknowingly. When you're talking about obviously prepubescent children, I'm in full agreement that it's almost always restricted to folks who seek it out explicitly.
Should we disallow encrypted traffic between machines because that encrypted traffic could be CSAM?
Saying "X is bad so we should be rid of it" requires a method to get rid of it.
Obviously giving up privacy and allowing more surveillance can reduce something but that isn't necessarily the right choice.
Like what do you do if they start sharing in files the user encrypts?
This is a very big assumption. Sexual abuse of minors has existed long before the internet, and long before photography. The notion that less availability of CSAM leads to less real-world abuse is not at all clear.
> If that's the best we can do, I'm very disappointed. That position says that to achieve privacy, I must tolerate CSAM. I want both privacy and for us not to tolerate CSAM. I don't know what the solution is, but that is what I wish the industry were aiming for. At the moment, the industry seems to be aiming for nothing but a shrug of the shoulders.
As other commenters have pointed out, the solution is to prevent children from being abused in the first place. Have robust systems in place to address abuse, and give kids effective education and somewhere to speak out if it happens to them or someone they know.
Yes, and you can tell because the proposed solutions attack privacy when alternative solutions exist.
For example, simply deleting CSAM material from devices locally without involving any other parties could have achieved the goals without privacy violations.
It makes me somewhat uncomfortable to argue for not involving other parties (like the police) in cases where real CSAM is found on someone’s device. Same as most people, I think that CSAM is morally reprehensible and really harmful to society. But just deleting it en-masse would have been an effective and privacy-respecting solution.
I think it’s important to see nuance even in things we don’t like to think about. Not everything that has a price tag has a price. We were told we needed to give up privacy, but that wasn’t necessary to take CSAM out of circulation.
...Understand I don't see it that way and applaud you for your way of thinking. I too have had to wrestle with the very uncomfortable "bed fellows" as it were that adhering to consistent application of principles inevitably results in.
The fact remains though that in a large swathe of the population, the ripping off and sacrifice of personal privacy is considered a small price to pay to inflict harm on that subpopulation. My issue comes in in that once you make the exception for one subpopulation, the slope is set.
Though even your "just delete it" has dystopian ramifications. Imagine that were implemented like Tianenmen Square? Recordings that are not blessed? You're still leaving in somebody's hands essentially executive control over what information can be allowed to exist, which is an unconscionably powerful lever to build.
This is one of those rare circumstances where "do nothing and clean up the mess" may be the most wise course of action.
Indeed, they are correct. And they were also brought up when Apple announced that they would introduce this Orwellian system. Now they act like they just realized this.
I think that companies might need to enable some kind of mechanism for offline investigation of the devices though. The CSAM is a real problem, there are real predators out there, the only risks isn't CSAM and law enforcement does really need to have a way to investigate devices. Previously, my proposal was the ability to force the device scan the user content for fingerprints of the suspected content but only with physical access. Physical access enforces the law enforcement to actually have a real and official investigation with strong enough reasons to spend resources and risk repercussions when done improperly.
However, the project of scanning all the user content for policing the users was one thing that irked and later relieved when Apple abandoned the project.
Apple's explanation is good and I agree with them but IMHO the more important aspects are:
1) Being able to trust your devices being on your side. That is, your device shouldn't be policing you and shouldn't be snitching on you. At this time you might think that the authorities who would have controlled yor device are on your side but don't forget that those authorities can change. Today the devices may be catching CSAM, some day the authorities can start demanding catching people opposing vaccines and an election or a revolution later they can start catching people who want to have an abortion or having premarital sexual relations or other non-kosher affairs.
2) Being free of the notion that you are always watched. If your device can choose to reveal your private thoughts or business, be it by mistake or by design, you can no longer have thoughts that are unaligned with the official ones. This is like the idea of a god who is always watching you but instead of a creator and angels you get C level businessmen and employees who go through your stuff when the devices triggers decryption of your data(by false positives or by true positives).
Anyway, policing everyone all the time must be an idea that is rejected by the free world, if the free world doesn't intent to be as free as Democratic People's Republic of Korea is democratic.
How many times will actual abusers be allowed to go free while your own family is victimized by the authorities and what ratio do you find acceptable?
Federal and state police, some of the best funded, equipped and trained police in the world are so inundated with cases that they are forced to limit their investigations to just toddlers and babies. What use is it to add more and more cases to a mountain of uninvestigated crimes? Whats needed is more police clearing the existing caseload.
That's an unbelievably depressingly low bar for society.
Oh, please. As if we couldn't just compare the hashes of the pictures people are storing against a CSAM database of hashes that gets regularly updated
When this was proposed people would respond "But they could just mirror the pictures or cut a pixel off!"
Who cares? You got that picture from some place in the dark web, and eventually someone will stumble upon it and add it to the database. Unless the person individually edits the pictures as they store them, that makes it so that you're never sure if your hashes will posthumously start matching against the DB.
People who wank off to CSAM have a user behavior similar to any other porn user, they don't store 1 picture, they store dozens, and just adding that step makes them likely to trip up, or straight up just use another service altogether
"What if there's a collision?" I don't know, go one step further with hashing a specific part of the file and see if it still matches?
This whole thing felt like an overblown fearmongering campaign from "freedom ain't free" individualists. I've never seen anything wrong with content hosters using a simple hash against you like this.
But even if there was an md5 hash collision back when md5 was the only one hash use, it still doesn't matter because upon viewing the image that matched, if it's not csam, it doesn't matter. Having said that, the chance of dozens of images matching hashes known to be associated to csam is also so unlikely as to be unthinkable. Where there is smoke, there is fire.
And further, a hash alone is meaningless, since in court there must be a presentation of evidence. If the image that set off the csam alarm by hash collision is say, an automobile, there is no case to be had. So all this talk about hash issues is absolutely moot.
Source: I have worked as an expert witness and presented for cases involving csam (back when we called it Child Pornography, because the CSAM moniker hadn't come about yet), so the requirements are well known to me.
Having said all that, I am an EFF member, and I prefer cryptography to work, and spying on users to be illegal.
That's why people were initially all up in arms about
But they didn't need to do that.
A regular hash where "just changing one pixel would be enough to evade the scanner" is already good enough
So when you say "a hash alone is meaningless, since in court there must be a presentation of evidence", you'd just present the image to court.
The hash is the trigger to call the authorities they handle the rest
And with a userbase the size of apple and people as pissy as reddit, you want to completely exclude a possibility of collisions or you'll get a repeat of the scenario we got in 2021
Blah blah blah, the same old argument given by the "think of the children" people.
There are many ways to counter that old chestnut, but really, we only need to remember the most basic fundamental facts:
1) Encryption is mathematics 2) Criminals are criminals
Can you ban mathematics ? No. Can you stop criminals being criminals ? No.
So, let's imagine you are able to successfully backdoor E2EE globally, on all devices and all platforms.
Sure, the "think of the children" people will rejoice and start singing "Hallelujah". And the governments will rub their hands with glee with all the new data they have access to.
But the criminals ? Do you honestly think they'll think "oh no, game over" ?
No of course not. They'll pay some cryptographer in need of some money to develop a new E2EE tool and carry on. Business as usual.
> No of course not. They'll pay some cryptographer in need of some money to develop a new E2EE tool and carry on. Business as usual.
I used to think this, I changed my mind: just as it's difficult to do security correctly even when it's a legal requirement, only the most competent criminal organisations will do this correctly.
Unfortunately, the other issue:
> And the governments will rub their hands with glee with all the new data they have access to.
Is 100% still the case, and almost impossible to get anyone to care about.
Assuming no new zero-days, and that they were even doing best practices in the first place. How many legit companies do best practices?
That's the point, bad actors can circumvent the system if they feel threatened, but system can be exploited by gov/companies once rolled out globally to target any user, so we get something that may be not that effective against bad actors but poses great risk to be misused by gov/company in their own interests without users knowing. I've seen how an authoritarian gov in my country is targeting ppl bc they are uncomfortable for the system and this algorithm opens another potential vector of attack.
> I've seen how an authoritarian gov in my country is targeting ppl bc they are uncomfortable for the system and this algorithm opens another potential vector of attack.
As per my last sentence in my initial comment in this chain:
--
> And the governments will rub their hands with glee with all the new data they have access to.
Is 100% still the case, and almost impossible to get anyone to care about.
All it takes is for one criminal to write a one-page guide to using GPG and circulate it to the group ....
I know I mentioned paying a cryptographer earlier, but in reality downloading and using GPG is a crude and effective way of defeating an E2EE backdoor.
Are the GPG devs going to backdoor GPG to satisfy governments ? Probably not.
If cybersecurity was that easy, we wouldn't have so many examples of businesses getting it wrong.
Just because everyone here can follow instructions like that, doesn't make it common knowledge for anyone else.
I can only partially agree with this point. Businesses getting cybersecurity wrong has almost no material and significant consequences. At best, they get a tiny slap on the wrist or asked to answer some questions. Nobody in said businesses goes to jail for it or personally pays any fines. Compare that to criminals who have a lot more to lose if they get caught — jail time, fines they have to pay, not having freedom for quite sometime, life not being the same after they’ve served their sentence, and more. Businesses have it extremely easy compared to this. No wonder cybersecurity is so poor among all businesses, including very large ones (like Microsoft, as a recent example).
Fear of these is the reason for the (maliciously compliant) GDPR popups, and that despite discussion about extra-territoriality and relativity limited capacity-to-websites ratio.
The law and threats of punishment are clearly not hugely significant to anyone involved in the specific topic of this thread regardless; in the UK at least, it's the kind of thing where if someone is lynched for it, the vigilantes have to be extremely stupid (like attacking a paediatrician because they can't tell the difference, which happened) to not get public sympathy.
There are almost no consequences for anyone working at a business that gets it wrong.
The consequences for being a nonce are quite severe so the motivation to get it right will be quite high.
GPG is infamous for being difficult to use correctly and for an antiquated design (IIRC forward secrecy is impossible?). And assuming E2EE backdoor actually exists, the gov is likely to be able to get at your key.
>Are the GPG devs going to backdoor GPG to satisfy governments?
No, but most users are unlikely to verify their GPG build is the right build.
[Citation needed]
This mindset—that assigns people into immutable categories, "criminal" and "not criminal"—is actually one of the biggest things that needs to change.
We absolutely can stop criminals from being criminals. We just can't do so by pointing at them and saying "Stop! Bad!" We have to change the incentives, remove the reasons they became criminal in the first place (usually poverty), and make it easier, safer, and more acceptable to go from being a criminal to being a not-criminal again.
Yet, the only justification around for mass surveillance for camera pictures, the important societal matter to rest literally on "the other side of the coin", is "some minority of people could be sharing naked photos of young members of society for highly unethical acts of viewing".
wtf.
https://www.nbcphiladelphia.com/news/local/child-porn-victim...
Absolutely.
> It is also bad for the fabric of society at large, in the sense that if we don't clearly take a stand against abhorrent behaviour then we are in some sense condoning it.
Much less clear. There's always been the argument: does it provide an outlet with no _new_ (emphasis on "new" so people don't skim that word) victims, or does it encourage people to act out their desires for real?
I don't have any answer to this question; but the answer matters. I do have a guess, which is "both at the same time in different people", because humans don't have one-size-fits-all responses.
Even beyond photographs, this was already a question with drawings; now we also have AI, creating new problems with both deepfakes of real people and ex-nihilo (victimless?) images.
> Does the tech industry have any alternate solutions that could functionally mitigate this abuse?
Yes.
We can build it into the display devices, or use a variation of Van Eck phreaking to the same effect.
We can modify WiFi to act as wall-penetrating radar with the capacity to infer pose, heart rate, and breathing of multiple people nearby even if they're next door, so that if they act out their desires beyond the screen, they'll be caught immediately.
We can put CCTV cameras everywhere, watch remotely what's on the screens, and also through a combination of eye tracking and (infrared or just noticing a change in geometry) who is aroused while looking at a forbidden subject and require such people to be on suppressants.
Note however that I have not said which acts or images: this is because the options are symmetrical under replacement for every other act and image, including (depending on the option) non-sexual ones.
There are places in the world where being gay has the death penalty. And if I remember my internet meme history right, whichever state Mr Hands was in, accidentally decriminalised his sex when the Federal courts decided states couldn't outlaw being gay and because that state had only one word in law for everything they deemed "unnatural".
I'd suggest there's a lot the not-tech industry could do to stop condoning abhorrent behavior that stops short of installing scanners on billions of computing devices. It's become a bit of a trope at this point, but it's bizarre to see a guy who is/was spokesman for a "minor attracted persons" (i.e. pedos) advocacy group getting published negatively reviewing the controversial new sex trafficking movie ... in Bloomberg:
https://www.bloomberg.com/opinion/articles/2023-07-15/qanon-...
For some background:
https://www.opindia.com/2021/08/meet-noah-berlatsky-prostasi...
His 501(c)(3) non-profit also advocates in favor of pedophilic dolls and has a "No Children Harmed certification seal" program for pedophilic dolls/etc:
https://prostasia.org/blog/dolls-prevent-child-sexual-abuse/
https://prostasia.org/no-children-harmed/
I'm not sure you can criminalize stuff like this, but it sets off my alarm bells when pedophile advocates are being published in mainstream news at the same time there's a moral panic around the need to scan everyone's hard drives. Is society actually trying to solve this problem, or is this more like renewing the Patriot Act to record every American's phone calls at the same time we're allied with al Qaeda offshoots in Syria? Interesting how terrorism has been the primary other argument for banning/backdooring all encryption.
----
As an aside I couldn't find ~anything about this group "Heat Initiative" Apple is responding to? Other than a TEDx talk by the founder a couple years ago which again seems very focused on "encrypted platforms" as the primary problem that needs solving: https://www.ted.com/talks/sarah_gardner_searching_for_a_chil...
That is a bigger problem and it will take a long time to fix. So long that I suspect that anybody reading this is long dead, but its like the saying with planting trees.
There is no parallel to be drawn between better encryption and worse outcomes for kids. Should we also outlaw high-performance cars because these sometimes serve as effective getaway vehicles for criminals?
CSAM producers and consumers should be found and punished via old-fashioned methods. How was this done in the past? Did we just never catch any human traffickers / rapists? No, we had detectives who went around detecting and presumably kicking down doors.
To outlaw large sections of mathematics because of this is absurd. And from the amount of power it would give big governments / big businesses, the fabric of society doesn't stand a chance.
Yea, LE/IC clearly have gone too far in many modern tactics.
Yea, it's possible to build a surveillance/police state much more efficiently than ever before.
Yea, we should be vigilant against authoritarianism.
What if we change the last bit after the "because" to "these sometimes are used at unsafe speeds and, intentionally or not, kill people who are not in cars?"
Because, at least for me, the answer is an unambiguous yes.
I agree that privacy and security should be available to everyone. But we also shouldn't count on being able to find people who are doing vile things--to children or adults--because the person messed up their opsec. I think Apple is correct here but as an industry we have to be putting our brains to thinking about this. "To outlaw large sections of mathematics" is hyperbole because we use mathematics to do a lot of things, some useful and some not.
How is this different from how police find anybody else who commits a crime? Like if they're trying to solve a murder, they're looking for DNA, clues, etc... They're literally looking for where the person who committed the crime "messed up their opsec" to use your wording.
Governments and law enforcement agencies have access to more information than they've ever had before. They have more cameras, location data, tons of data compiled by data brokers. But it's not enough - of course they have to have this too.
On top of that, there's a long history now of governments buying zero day vulnerabilities or even technology from firms like NSO and guess what? It's not being used to catch pedophiles (cue shocked Pikachu face) but it is being used to target political dissidents.
This is so frustrating, because it feels like a siege on a city. Collectively, people have to fight against bad legislation in various countries constantly. But the other side only has to "win" once.
I'd say that "we" as an industry should be putting out brains to trying to figure out ways to make it even harder for these people to legislate encryption out of existence, than trying to find ways to appease geriatric lawmakers.
Recently invented encrypted chat rooms allow people to coordinate and transfer CSAM without any government official being able to infiltrate it. And just being able to freely discuss has been shown to make the problem worse as it facilitates knowledge transfer.
This is all completely different to in the past where this would have been done in person. So the argument that we should just do what we did in the past makes no sense. As technology advances we need to develop new techniques in order to keep up.
Any source for that statement ? My understanding is that child abuse - while still existing - is at an all time low in modern western societies. Children working in brothel used to be common, children getting abused by the church used to be a well know fact etc.
Overall a using children is harder than ever even if it has not completely disappeared.
Technology - which allows efficient transfer of information, testimonies etc - is instrumental to that evolution.
The "old-fashioned methods" that they used in the past included intercepting communications of people that were suspected of crimes, such as by getting a warrant allowing them to force the person's phone company to record and turn over the person's calls, or by getting a warrant to intercept and inspect the contents of the person's mail at the post office.
> To outlaw large sections of mathematics because of this is absurd
No one has or is proposing outlawing large sections of mathematics, or even small sections of mathematics. The laws are outlawing some applications that make use of mathematics.
Calling that outlawing mathematics is as absurd as saying that building codes that won't let me use asbestos insulation in new construction are banning sections of thermodynamics. Or saying that laws that restrict how high I can fly a drone are banning large sections of aerodynamics.
Stop equivocating. You're banning the mathematics. The mechanism is literally the mechanical implementation of the mathematics.
>Calling that outlawing mathematics is as absurd as saying that building codes that won't let me use asbestos insulation in new construction are banning sections of thermodynamics.
...Except that's not even an analogous comparison? The asbestos is forbidden not because it's too good an insulator/foiler of thermodynamics, but because of it's danger to the health of everyone.
Trying to ban applications that use encryption is exactly banning asbestos because it's too good an insulator, and you're interested in seeing whatever is wrapped in it burn.
Yes. If consumer cars’ speed was capped to 100 mph it would affect a very small percentage of people and zero legitimate use.
The difference with encryption is that while it protects criminals from the police, it also protects legitimate users from criminals.
How many more times do we have to see this movie before people remember how it ends.
I don't really buy any "slippery slope" arguments for this stuff. Apple already can push any conceivable software it wants to all of its phones, so the slope is already as slippery as it can possibly be.
It just doesn't make sense to say "Apple shouldn't implement this minimal version of photo-scanning now even though I don't think it's bad, because that's a slippery slope for them to implement some future version of scanning that I do think is bad." They already have the capability to push any software to their phones at any time! They could just skip directly to the version you think is bad!
your source for proof of that?
>> in the sense that if we don't clearly take a stand against abhorrent behaviour then we are in some sense condoning it.
No. this narrative of "silence is violence" and "no action is support" etc is 100% wrong.
You started out great, but i can not get behind this type of thinking...
>>Does the tech industry have any alternate solutions that could functionally mitigate this abuse?
Why is it a "tech industry" problem
>>Or do we all just shout "yay, individual freedom wins again!"
For me the answer is simple... Yes individual freedom is more important that everything. I will never support curbing individual freedom on the alter of any kind proclaimed government solution to a social safety problem. Largely because I know enough about history to understand that not only will they no solve that social safety problem, many in government are probably participating in the problem and have the power to exempt themselves, while abusing the very tools and powers we give them to fight X, for completely unrelated purposes.
Very quickly any tool we would give them to fight CSAM would be used for Drug Enforcement, Terrorism, etc. It would not be long before the AI based phashes detect some old lady's Tomato plants as weed and we have an entire DEA paramilitary unit raiding her home...
Creating backdoors that allow encryption schemes to be subverted is _fundamentally_ going to cause harm on the internet, and eventually fail the weakest users/those that need privacy/security the most.
A mechanism that can subvert cryptographic protocols can be used by any party, including oppressive regimes, private entities etc. that have the resources/will/knowledge to use the backdoor etc. Backdoors harm both the trust on the web (which can have an impact on economic transactions among many others) and the people that need security/privacy the most. In the meantime, criminals will wise up and move their operations elsewhere where no backdoors exist.
We basically end up with a broken internet, we are putting people in harm's way and the criminals we are targeting are probably updating their OPSEC/MO not to rely on E2EE.
...regardless of whether Apple rolls out E2EE right? End to end encryption is available through a whole host of open-source tools, and should Apple deploy CSAM scanning the crooks will just migrate to a different chat tool.
A small percentage of people are involved in this crime and subjecting every single person to illegal searches and possibly getting wrongly identified or nation states using this to imprison its enemies is wrong.
Sometimes there is no solution that satisfies everyone and sometimes the only good solution is the least shittiest solution but still a shitty solution. In this case, I believe that my and everyone else’s freedoms and privacy are worth it and we should instead spend much money time and effort trying to catch these criminals instead of scanning everyone’s phones which won’t ultimately work.
You mean at rates greater than in the 1970s when it was legal in the US and there were a bunch of commercial publications creating and distributing it (prior to the "Protection of Children Against Sexual Exploitation Act" in 1977 (which targeted the production) and "Child Protection and Obscenity Enforcement Act" in 1988 (which targeted the distribution))? I find that doubtful.
> Does the industry feel that it has any responsibility at all to do so?
Why should it? Should every home builder and landlord install a camera in every bedroom to stream back to the government to be sure that no children are being molested? Why not? It's technologically possible today.
The fact that garbage people do garbage things to vulnerable people doesn't mean that the entire world must be commandeered to stop it.
E2E encryption has replaced a lot of in person meeting that also would have been secure against dragnet surveillance. We shouldn't lose our privacy just because it became technically possible to take it away, no different than the cameras in bedrooms mentioned above.
> a new child safety group known as Heat Initiative
Doesn't even have a website or any kind of social media presence; it literally doesn't appear to exist apart from the reporting on Apple's response to them, which is entirely based on Apple sharing their response with media, not the group interacting with media.
> Sarah Gardner
on the other hand previously appeared as the VP of External Affairs (i.e. Marketing) of Thorn (formerly DNA Foundation): https://www.thorn.org/blog/searching-for-a-child-in-a-privat...
So despite looking a bit fishy at first, this doesn't seem to come from a christofascist group.
Why would you assume this in the first place?
I mean, there aren't many people who want paedophiles to be able to amass huge collections of child abuse imagery from other paedophiles online. And pretty much every parent wants their child to be kept safe from predators both online and offline.
And yes, because their politics are becoming normalized within American culture, it is necessary to be skeptical about references to any such group. Assuming good faith is a rule on HN but elsewhere, where bad faith is what gets visibility, it's naive.
https://www.theguardian.com/politics/2014/mar/02/how-paedoph...
So if there are indeed some right-wing groups talking about this, maybe it's best not to brush off their claims without some scrutiny first. And I say this as someone who mostly agrees with the left on most things.
Anyway I don't think that any of this has much to do with Apple being asked to implement specific technical measures for detecting child abuse imagery.
Figlet indeed.
> figlett 5 months ago [flagged] [dead] | parent | context | prev [–] | on: Florida courts could take 'emergency' custody of k...
> This is excellent news for children at risk of being abused by militant transgenders and the medical establishment who are enabling them. Thank you Florida for trying to put an end to this menace.
I'm in the UK and we're doing better here though, the main left-wing party is backing away from the particular ideology that has enabled this. I was going to vote for them anyway as we desperately need our public services to be restored and welfare for those less fortunate in society to be improved, but I'm pleased they're moving towards a sensible, harm-reducing stance on this issue rather than assuming everything the gender activists say is reasonable.
So yeah. When these things pop up I assume malicious intent.
> Pornography is humiliation and degradation of women. It's a disgraceful activity. I don't want to be associated with it. Just take a look at the pictures. I mean, women are degraded as vulgar sex objects. That's not what human beings are. I don't even see anything to discuss.
> Interviewer: But didn't performers choose to do the job and get paid?
> The fact that people agree to it and are paid, is about as convincing as the fact that we should be in favour of sweatshops in China, where women are locked into a factory and work fifteen hours a day, and then the factory burns down and they all die. Yeah, they were paid and they consented, but it doesn't make me in favour of it, so that argument we can't even talk about.
> As for the fact that it's some people's erotica, well you know that's their problem, doesn't mean I have to contribute to it. If they get enjoyment out of humiliation of women, they have a problem, but it's nothing I want to contribute to.
> Interviewer: How should we improve the production conditions of pornography?
> By eliminating degradation of women, that would improve it. Just like child abuse, you don't want to make it better child abuse, you want to stop child abuse.
> Suppose there's a starving child in the slums, and you say "well, I'll give you food if you'll let me abuse you." Suppose - well, there happen to be laws against child abuse, fortunately - but suppose someone were to give you an argument. Well, you know, after all a child's starving otherwise, so you're taking away their chance to get some food if you ban abuse. I mean, is that an argument?
> The answer to that is stop the conditions in which the child is starving, and the same is true here. Eliminate the conditions in which women can't get decent jobs, not permit abusive and destructive behaviour.
(Source of the above is this interview: https://youtube.com/watch?v=SNlRoaFTHuE)
You can rest assured that they are scanning all of it serverside for illegal images presently.
The kerfuffle was around clientside scanning, something that it has been reported that they dropped. I have thus far seen no statements from Apple that they actually intended to stop the deployment of clientside scanning.
Serverside scanning has been possible (and likely) for a long time, which illuminates their "slippery slope" argument as farce (unless they intend to force migrate everyone to e2ee storage in the future).
Approximately no one uses it.
Hopefully Apple will begin promoting users to migrate in future updates.
Mind sharing where it is on an iPhone and Mac? I have not been able to find it.
It's kinda complicated to turn on, as it disables most account recovery options.
Anyone who doesn't know about this feature is not a power user in my book.
Whatever they said, it was probably worded to give you this impression, without actually saying that. Apple is extremely careful and goes to great pains to actively mislead and deceive whilst avoiding actual lies.
Could you please link to or quote these statements from Apple? I would bet any money they say something different than what you claim, a "not wittingly"-style hedge.
Apple clearly has very limited ongoing scanning because they report on the order of hundreds of instances of CSAM to NCMEC every year whereas other services with pervasive scanning report on the order of tens of millions of instances.
It’s ok, you are one of the vast majority of people commenting on this topic while reasoning from false premises.
> Then how would they respond to warrants asking for all user identities that upload image x?
Dragnet warrants like this aren’t even legally permissible in the United States, and if they were Apple would obviously reject them. They state that they only provide specific named user account information in response to warrants.
The headline says they don't scan iCloud Photos, but I don't see the statements from Apple saying that. The media often misreports on Apple's comments because Apple's expert at inducing the media to misreport in ways that are favorable to Apple.
The earlier design was a hybrid model that scanned for CSAM on device, then flagged files were reviewed on upload.
Note that by design the hashes cannot be audited (though in the legitimate case I don’t imagine doing so would be pleasant), so there’s nothing stopping a malicious party inserting hashes of anything they want - and then the news report will be “person x bought in for questioning after CSAM detector flagged them”.
That’s before countries just pass explicit laws saying that the filter must includE LGBT content (in the US several states consider books with lgbt characters to be sexual content, so a lgbt teenager would be de facto CSAM), in the UK the IPA is used to catch people not collecting dog poop so trusting them not to expand scope is laughable, in Iran a picture of a woman without a hijab would obviously be reportable, etc
What Apple has done is add the ability to filter content (eg block dick picks) and for child accounts to place extra steps (incl providing contact numbers I think?) if a child attempts to send pics with nudity, etc
What does this mean? What is IPA? I tried Googling for it but I’m not finding much. I would love to learn more about that
It was passed to stop terrorism, because previously they found that having multiple people (friends and family etc) report that someone was planning a terrorist attack failed to stop a terrorist attack.
But you're missing the point:
Step 1. generate some opaque hash of the "semantics" of an image
Step 2. compare those hashes to some list of hashes of "CSAM", which again fundamentally cannot be audited
Step 3. report any hits to law enforcement
Step 4. person X is being investigated due to reported violations of laws against child abuse.
Basically: how do you design a system in which the state provides "semantic" hashes of "CSAM" that cannot be trivially abused by inclusion of non-CSAM as "CSAM", or by laws mandating inclusion of things that are objectively not-CSAM. Hypothetically: hashes that match christian crosses, star of David, muslim star and/or crescent, etc. Or in the US DNC, RNC, pride, etc flags. Recall that definitionally no one can audit the hashes that would trigger notifying law encforcement.
In a word, decentralization.
By detecting unsafe material on-device / while it is being created, they can prevent it from being shared. And because this happens on individual devices, Apple doesn’t need to know what’s on people’s iCloud. So they can offer end-to-end encryption, where even the data on their servers is encrypted. Only your devices can “see” it (it’s a black box for Apple servers, gibberish - without the correct decryption key).
It is amazing that so much counter-cultural spirit remains in Apple. They are probably going to ban likes and other vanity features in all iOS applications, prohibit access to popular media, put “pop stars” into rehabs, and teach their users to disobey (the hardest of all tasks).
A lot of people try really hard not to see that “unusual” abuse of the children is the same as “usual” abuse of everyone. Conveniently, the need for distinction creates “maniacs” that are totally, totally different from “normal people”, and cranks up the sensation level. The discussion of “external” evil then can continue ad infinitum without dealing with status quo of “peaceful, normal life”.