News for hackers. Stuff that matters.
Re:Hash collision in Apple NeuralHash model by Grustaf on Wednesday August 18, 2021 @09:27AM (#28220673)
Why would they extend the CSAM scanner? It would be much simpler to just use all the OCR and image classification functions they have already deployed.

CSAM scanning is only useful for areas where Apple really doesn't want to even look at the actual material until they are extremely certain that it's a match. If they want to detect anti-government propaganda or something, there would be no such concerns, they would just do a regex search of your inbox. Infinitely more convenient.

Re:Hash collision in Apple NeuralHash model by sandworm101 on Wednesday August 18, 2021 @09:43AM (#28220828)
>> useful for areas where Apple really doesn't want to even look at the actual material

Correct. It has plausible deniability built in. Apple is unable to verify that the images the government are looking for are actually CSAM. They could be political. They could be protest images. They could be Winnie the Pooh. Apple can plead ignorance as it blindly scans for whatever the requesting government asks it to scan for.

Nobody really minds that this system is going to be used for CSAM. What everyone recognizes is how ripe this system is for abuse, how easily it can be leveraged by oppressive governments. And Apple can play the innocent.

Re:Hash collision in Apple NeuralHash model by mlindner on Wednesday August 18, 2021 @10:11AM (#28221123)
> Nobody really minds that this system is going to be used for CSAM.

I beg to differ. It doesn't matter how evil the content is, no scanning of my computers by outside parties, period. More so by scanning law enforcement can even plant legitimate child pornography on people's computers and get convictions all the easier because the system self-reports.

Re:Hash collision in Apple NeuralHash model by Grustaf on Wednesday August 18, 2021 @10:18AM (#28221251)
It's clearly not the scanning you are opposed to, it's the reporting. Apple apps, well a lot of apps, already scan all your content, for classifying images, doing OCR, extracting dates and addresses etc. That is nothing new, it's omnipresent.
Re:Hash collision in Apple NeuralHash model by jasamer on Wednesday August 18, 2021 @12:34PM (#28223632)
What about the human reviewers at Apple? Those need to confirm matches, and they obviously look at the photos to do so. Apple can’t claim ignorance.

Also, it has to be two requesting governments.

Re:Hash collision in Apple NeuralHash model by sandworm101 on Wednesday August 18, 2021 @12:49PM (#28223851)
The current Apple policy to use human reviewers. That might change at a moment's notice and there is no technical reason why Apple couldn't bypass humans for certain requests. One must always judge a new system three ways: how it was meant to be implemented, how it is actually implemented, and how it might be abused by bad actors in the future.
Re:Hash collision in Apple NeuralHash model by Grustaf on Wednesday August 18, 2021 @10:00AM (#28221006)
Except that Apple will review the photos once you've matched 30 of them, so it's still not possible for the government to misuse it.
Re:Hash collision in Apple NeuralHash model by short_sells_poo on Wednesday August 18, 2021 @10:47AM (#28221716)
So some underpaid contractor reviewing "visual derivatives" that may or may not be CSAM completely prevents governments from misusing this in your mind?
Re:Hash collision in Apple NeuralHash model by Grustaf on Wednesday August 18, 2021 @12:18PM (#28223423)
Considering how high profile and incredibly sensitive this is, I doubt they will hire an underpaid contractor. It's the opposite of app review, where the volume is very high but the stakes are low.
Re:Hash collision in Apple NeuralHash model by MichaelGroves on Wednesday August 18, 2021 @12:46PM (#28223810)
Facebook hires underpaid and poorly treated contractors to moderate exactly this same kind of content. I see no reason to believe Apple will be any different, particularly a few months from now when the general public's attention has shifted to other matters (assuming they're even paying attention right now. I don't think the interests of HN are necessarily representative of the general public..)
Re:Hash collision in Apple NeuralHash model by kps on Wednesday August 18, 2021 @10:10AM (#28221105)
In China, iCloud is run by the government.
Re:Hash collision in Apple NeuralHash model by Grustaf on Wednesday August 18, 2021 @10:16AM (#28221210)
It's actually not, but even if it were, that would be yet another reason CSAM scanning is completely irrelevant for government spying.

Any spying really. It's much easier to just look at the images themselves.

Re:Hash collision in Apple NeuralHash model by ttflee on Wednesday August 18, 2021 @11:29AM (#28222564)
> It's actually not

It's Joe Wong's joke, that it's like peeing in the snow in a dark winter night, while there is a difference but it's really hard to tell.

Re:Hash collision in Apple NeuralHash model by ethbr0 on Wednesday August 18, 2021 @09:49AM (#28220892)
Why even keep Apple in the loop? Why not just allow government to submit scanning models directly?

Which Apple will dutifully install and run, because they're required by local laws.

Re:Hash collision in Apple NeuralHash model by jodrellblank on Wednesday August 18, 2021 @10:18AM (#28221253)
> "Which Apple will dutifully install and run, because they're required by local laws."

Which Apple have stated that they won't do, and have designed the system so they can't do that without it being found out: https://news.ycombinator.com/item?id=28221082

Can't you at least post accurate information about this system and support outrage based on facts instead of fantasy?

Re:Hash collision in Apple NeuralHash model by mschuster91 on Wednesday August 18, 2021 @11:10AM (#28222149)
> Which Apple have stated that they won't do

For your random off-of-the-mill dictatorship, yes.

For the US? EU? China? India? No way they can refuse such a request from these markets. And if they could and get away with it, it would be a very worrying situation in itself regarding (democratic) control of government over global mega corporations.

Re:Hash collision in Apple NeuralHash model by jodrellblank on Wednesday August 18, 2021 @11:28AM (#28222545)
If you believe "the government can compel any company to do anything", then this system is no worse than any other.

Although how do you think Linus Torvalds "manged to get away with refusing" adding backdoors? https://www.techdirt.com/articles/20130919/07485524578/linus...

Re:Hash collision in Apple NeuralHash model by mschuster91 on Wednesday August 18, 2021 @12:58PM (#28224001)
Simple reason: every line of code in the Linux kernel passes many eyeballs in the open. Even if someone were to compel Linus Torvalds to commit a backdoor and he would comply with that order, people across many jurisdictions would notice this immediately, rendering the backdoor worthless.