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

The Problem with Perceptual Hashes

rentafounder.com

Post reply on HN
Print view View original

171–180 of 440 posts

Re: The Problem with Perceptual Hashes

#171

Post by axaxs »

yellow_lead wrote: post #61

Regarding false positives re:Apple, the Ars Technica article claims > Apple offers technical details, claims 1-in-1 trillion chance of false positives. There are two ways to read this, but I'm assuming it means, for each scan, there is a 1-in-1 trillion chance of a false positive. Apple has over 1 billion devices. Assuming ten scans per device per day, you would reach one trillion scans in ~100 days. Okay, but not al…

Eh...I don't think of it as one in a trillion scans...but one in a trillion chance per image. I have something like 2000 pics. My wife, at least 5x that number. If we split the difference, and assume the average device has 5000 pics, that's already hitting false positives multiple times. Feel sorry for the first 5 to get their account banned on day 1 because their pic of an odd piece of toast was reported to the govt as cp.

Re: The Problem with Perceptual Hashes

#172

Post by dannyw »

gpm wrote: post #156
mrtksn wrote: post #72

The technical challenges aside, I’m very disturbed that my device will be reporting me to the authorities. That’s very different from authorities taking a sneak peek into my stuff. That’s like the theological concept of always being watched. It starts with child pornography but the technology is indifferent towards it, it can be anything. It’s always about the children because we all want to save the children. Soon t…

With you up to here, but this is jumping the shark > I bet you, after the next election in the US your device will be reporting you for spreading far right or deep state lies, depending on who wins. The US is becoming less stable, sure [1], but there is still a very strong culture of free speech, particularly political speech. I put the odds that your device will be reporting on that within 4 years as approximately 0…

Did you literally not see a former president effectively get silenced in the public sphere by 3 corporations?

How can you seriously believe that these corporations (who are not subject to the first amendment, and cannot be challenged in court) won't extend and abuse this technology to tackle "domestic extremism" but broadly covering political views?

Re: The Problem with Perceptual Hashes

#173

Post by wyager »

gpm wrote: post #169
efitz wrote: post #160

Earlier quoted context omitted.

Apple has a shitty record wrt free speech. Apple hates free speech. Apple likes “curation”. They canned Parler in a heartbeat; they also police the App Store for anything naughty.

Canning Parler is Apple choosing not to advertise and send you an app they don't like, i.e. it's Apple exercising it's own right to free speech. Agree or disagree with it, it's categorically different from Apple spying on what the files you have are saying (not even to or via Apple) and reporting it to the government.

Apple also disallows you from installing things without going through them, so “choosing not to advertise and send” has a lot more significance than your wording implies.

It’s not like they have a curated App Store for apps they like; there’s literally no other way to add software to the device.

Re: The Problem with Perceptual Hashes

#174

Post by bastawhiz »

Correct me if I'm wrong, but nowhere in Apple's announcement do they mention "perceptual" hashing. I've searched through some of the PDFs they link as well, but those also don't seem to mention the word "perceptual". Can someone point out exactly where this is mentioned?

Re: The Problem with Perceptual Hashes

#175

Post by dannyw »

wonnage wrote: post #150
Renaud wrote: post #144

Earlier quoted context omitted.

What has changed is the inclusion of spyware technology on the device that can be weaponised to basically report on anything. Today it's only geared toward iCloud and CSAM. How many lines of codes do you think it will take before it scans all your local pictures? How hard do you think it will be for an authoritarian regime like China, that Apple bends over backwards to please, to start including other hashes that are…

We gotta stop with the China bogeyman every time a privacy issue comes up. This is a feature designed by an American company for American government surveillance purposes. China is perfectly capable of doing the same surveillance or worse on its own citizens, with or without Apple. China has nothing to do with why American tech is progressively implementing more authoritarian features in a supposedly democratic count…

China is just an example. In Australia, the law allows our executive department to order tech companies to build backdoors for the investigation of any crime punishable by more than 2 years imprisonment.

We actually had the anti terror department arrest a popular, left-leaning YouTube influencer for harassment while physically assaulting his mum (all on video).

That's something that is literally unprecedented in Hong Kong just 3 years ago.

Re: The Problem with Perceptual Hashes

#176

Post by dannyw »

Klonoar wrote: post #137
stetrain wrote: post #131

Earlier quoted context omitted.

The post office scans your mail through various machines in transit. We accept that when we put the mail in the mailbox. What if the post office announced they were installing a man with a scanning machine in your home who would scan your letters before they left your house? It's the same outcome. The same process. Just inside your house instead of out in the mail system. They're exactly the same, except somehow it's…

This example changes with regards to emotional weight if you remove "a man" and leave it at just "a scanning machine". There is no human scanning your photos on an iPhone, so let's compares apples to apples here. If that scanning machine didn't reveal the contents of my mail, and then ensured that it wasn't able to be given out in-transit? Yeah, I'd potentially be fine with it - but I'll leave this answer as a hypoth…

There actually is a man involved: enough similarities and a human will review the photos. Every algorithm, especially perceptual hashing, will have false positives, and at Apple's scale, some people's private and intimate photos will be false positives and be exposed to a man looking at it.

Re: The Problem with Perceptual Hashes

#177

Post by ssklash »

dylan604 wrote: post #138
mrtksn wrote: post #72

The technical challenges aside, I’m very disturbed that my device will be reporting me to the authorities. That’s very different from authorities taking a sneak peek into my stuff. That’s like the theological concept of always being watched. It starts with child pornography but the technology is indifferent towards it, it can be anything. It’s always about the children because we all want to save the children. Soon t…

>but I’m not going to carry a snitch in my pocket. I wonder how this will hold up against 5th ammendment (in the US) covering self-incrimination?

I assume the third party doctrine makes it so that the 5th amendment doesn't apply here.

Re: The Problem with Perceptual Hashes

#178

Post by dannyw »

fogof wrote: post #155
ezoe wrote: post #110

The problem of hash or NN based matching is, the authority can avoid explaining the mismatch. Suppose the authority want to false-arrest you. They prepare a hash that matches to an innocent image they knew the target has in his Apple product. They hand that hash to the Apple, claiming it's a hash from a child abuse image and demand privacy-invasive searching for the greater good. Then, Apple report you have a file th…

Well, presumably at that point, someone in that position would just reveal their own files with the hash an prove to the public that they weren't illegal. Sure, it would be shitty to be forced to reveal your private information that way, but you would expose a government agency as fabricating evidence and lying about the contents of the picture in question to falsely accuse someone. It seems like that would be a scan…

There are literally hundreds of cases of police fabricating evidence and getting caught in court, or on bodycam.

This happens today. We must not build technology that makes it even more devastating.

Re: The Problem with Perceptual Hashes

#179

Post by rcarback »

bastawhiz wrote: post #174

Correct me if I'm wrong, but nowhere in Apple's announcement do they mention "perceptual" hashing. I've searched through some of the PDFs they link as well, but those also don't seem to mention the word "perceptual". Can someone point out exactly where this is mentioned?

"NeuralHash is a perceptual hashing function"

https://www.apple.com/child-safety/pdf/CSAM_Detection_Techni...

Re: The Problem with Perceptual Hashes

#180

Post by wilg »

yellow_lead wrote: post #61

Regarding false positives re:Apple, the Ars Technica article claims > Apple offers technical details, claims 1-in-1 trillion chance of false positives. There are two ways to read this, but I'm assuming it means, for each scan, there is a 1-in-1 trillion chance of a false positive. Apple has over 1 billion devices. Assuming ten scans per device per day, you would reach one trillion scans in ~100 days. Okay, but not al…

Apple claims that metric for a false positive account flagging, not photo matching.

> The threshold is set to provide an extremely high level of accuracy and ensures less than a one in one trillion chance per year of incorrectly flagging a given account.

https://www.apple.com/child-safety/

Post reply on HN