Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Apple defends anti-child abuse imagery tech after claims of ‘hash collisions’

vice.com

Post reply on HN
Print view View original

41–50 of 465 posts

Re: Apple defends anti-child abuse imagery tech after claims of ‘hash collisions’

#41

Post by robertoandred »

geoah wrote: post #5

> The system relies on a database of hashes—cryptographic representations of images—of known CSAM photos provided by National Center for Missing & Exploited Children (NCMEC) and other child protection organizations. “Cryptographic representations of images”. That’s not the case though right? These are “neuralhashes” afaik which are nowhere close to cryptographic hashes but rather locality sensitive hashes which is a…

The CSAM hashes are encrypted and blinded, possible Vice and others conflated and/or combined steps.

Re: Apple defends anti-child abuse imagery tech after claims of ‘hash collisions’

#42

Post by epistasis »

What's shocking to me is how little Apple management understood of what their actions looked like. Really stunning.

For a company that marketed itself as one of the few digital service providers that consumers could trust, I just don't understand how they acted this way at all.

Either there will be heads rolling at management, or Apple takes a permanent hit to consumer trust.

Re: Apple defends anti-child abuse imagery tech after claims of ‘hash collisions’

#43

Post by Barrin92 »

it's kind of silly how this is treated like some deeply technical issue. Apple's 'one of a trillion' claim is either true and the software is useless, because I'm pretty sure pedophiles can figure out what a watermark or a gaussian blur in paint net is, or it's imprecise and actually detects things which is the very thing that makes it dangerous to the public.

It's a direct trade-off and the error tolerance of any such filter is the only thing that makes it useful so we can basically stop arguing about the depths of implementation details or how high the collision rate of the hashing algorithm is etc. If this thing is supposed to catch anyone it needs to be magnitudes more lenient than any of those minor faults.

Re: Apple defends anti-child abuse imagery tech after claims of ‘hash collisions’

#44

Post by _zzaw »

firebaze wrote: post #31

I think we're up for a new Apple CEO in a few days.

The turmoil of Apple abruptly booting the CEO who's overseen the most obscenely prosperous stretch in its history would cost the company vastly more than the current controversy about CSAM.

I'm not saying this as a fan of either Cook or their anti-CSAM measures; I'm neither, and if anyone is ever wrongfully arrested because Apple's system made a mistake, Cook may well wind up in disgrace depending on how much blame he can/can't shuffle off to subordinates. I don't think we're there yet, though.

Re: Apple defends anti-child abuse imagery tech after claims of ‘hash collisions’

#45

Post by justapassenger »

epistasis wrote: post #42

What's shocking to me is how little Apple management understood of what their actions looked like. Really stunning. For a company that marketed itself as one of the few digital service providers that consumers could trust, I just don't understand how they acted this way at all. Either there will be heads rolling at management, or Apple takes a permanent hit to consumer trust.

> For a company that marketed itself as one of the few digital service providers that consumers could trust, I just don't understand how they acted this way at all.

Because privacy stance is mostly PR to differentiate from Google. And while there're invalid reasons to get users data, there're also valid ones (at least from legal requirement point of view - let's not get into weeds about personal freedom here and if the laws and its implementations need to be changed).

Their PR was just writing the checks they cannot cash without going on a war with governments.

Re: Apple defends anti-child abuse imagery tech after claims of ‘hash collisions’

#46

Post by rootusrootus »

whoknowswhat11 wrote: post #30
geoah wrote: post #5

> The system relies on a database of hashes—cryptographic representations of images—of known CSAM photos provided by National Center for Missing & Exploited Children (NCMEC) and other child protection organizations. “Cryptographic representations of images”. That’s not the case though right? These are “neuralhashes” afaik which are nowhere close to cryptographic hashes but rather locality sensitive hashes which is a…

No - the reporting is absolutely terrible here. 1) These are more share similar visual features than crypto hashes. 2) HN posters have been claiming that apple reviewing flagged photos is a felony -> because HN commentators are claiming flagged photos are somehow "known" CASM - this is also likely totally false. The images may not be CASM and the idea that a moderation queue results in felony charges is near ridiculo…

> The images may not be CASM and the idea that a moderation queue results in felony charges is near ridiculous.

Agree, and I think this is backed up by real world experience. Has Facebook or anyone working on their behalf ever been charged for possession of CSAM? I guarantee they've seen some. Probably a lot, in fact. That's why we have recurring discussions about the workers and the compensation they get (or not) for the really horrid work they are tasked with.

Re: Apple defends anti-child abuse imagery tech after claims of ‘hash collisions’

#47

Post by daxuak »

epistasis wrote: post #42

What's shocking to me is how little Apple management understood of what their actions looked like. Really stunning. For a company that marketed itself as one of the few digital service providers that consumers could trust, I just don't understand how they acted this way at all. Either there will be heads rolling at management, or Apple takes a permanent hit to consumer trust.

They are very likely aware of backlash but since this is a easily defendable hill with a very slippery slope down the way, it is of their interest to push for it IMHO.

Re: Apple defends anti-child abuse imagery tech after claims of ‘hash collisions’

#48

Post by jedmeyers »

gjsman-1000 wrote: post #4

For everyone upset about Apple's CSAM scanning, I think we all forgot about the EARN IT Act. It was nearly passed last year but Congress was finished before it could be voted on. It had Bipartisan support and would've virtually banned E2E of any kind. And it would have required scanning everywhere according to the recommendations of a 19-member board of NGOs and unelected experts. The reason for this mandatory backdo…

> If you are Apple, even though EARN IT failed... you know where Washington's heart lies.

If you are IG Farben, you know where Berlin's heart lies...

Re: Apple defends anti-child abuse imagery tech after claims of ‘hash collisions’

#50

Post by yeldarb »

salawat wrote: post #34
yeldarb wrote: post #15

Earlier quoted context omitted.

Doing that would mean the NeuralHash would change though. And you'd have to not only get CLIP to identify CSAM in the generated image but also negate the parts of the generated image that are causing CLIP to label it as "generated" (while still colliding with the target NeuralHash). Unclear how hard this would actually be in practice (if I were going to attempt it, the first thing I'd try is to evolve a colliding ima…

Take 2 CSAM pictures, known. Combine into one image. Swing and a miss. Not in the CSAM dataset. Take two images. Encode alternating pixels. Decode to get the original image back. Convert to different encodings print to PDF or Postscript. Encode as base64 representations of the image file... Who are we trying to fool here? This is kiddie stuff. This is more about trying to implant scanning capabilities on client devic…

Those are examples of triggering false negatives (CSAM that gets missed by the system), this collision attack is about triggering false positives (non-CSAM that triggers human intervention).
Post reply on HN