4 ms·

Would this work as an attack? 1. Get a pornographic picture involving young though legal actors and actresses. 2. Encode a nonce into the image. Hash it chec

Would this work as an attack? 1. Get a pornographic picture involving young though legal actors and actresses. 2. Encode a nonce into the image. Hash it checking for CSAM collisions. If you've found a collision go on to the next step, if not update the nonce and try again. 3. You now have an image that, to visual inspection will appear plausibly like CSAM, and to automated detection will appear like CSAM. Though, presumably, it is not illegal for you to have this image as it is, in fact, legal pornography. You can now text this to anyone with an iPhone who will be referred by Apple to law enforcement.

arn 5y ago
How would you know if it's a CSAM collision? I don't believe that database is publicly available anywhere, for obvious reasons.
- arsome 5y ago
  Apple's client clearly has some revealing information on that...
- ALittleLight 5y ago
  Good point. I can offer three possibilities to how you might know. First, a data leak of the material. Second, law enforcement presumably gives hashes of known CSAM to service operators so they can detect and report it. You could pose as or be the operator of such a service and get the hashes that way. Third, if you were a government operator you may have access to the hashes that way. (Although I guess corrupt government agents would have other easier ways of getting to you)
- FabHK 5y ago
  Yes. That, plus it's unclear whether you can create a collision with a nonce. It's a perceptual hash, not a cryptographic one. Lastly, to compute that collision might be so expensive that you could instead compute small SHA-256 hashes, ie mine BTC, and use the money to obtain your sinister goals via other means.
Ashanmaril 5y ago
If you're gonna go through that much effort, you might as well just text someone an actually illegal image from something other than an iPhone You're already attempting to frame someone for a crime, might as well commit another crime while you're at it
robertoandred 5y ago
Why would you save to your personal photo library some porn that a creepy rando sent to you in a text?
shapefrog 5y ago
Your hash will match to say image 105 of the dataset. Upon visual inspection, your 'legal porn' is going to have to at least look passably like image 105 of the dataset to get anywhere. So at this point we have an image that computers think is CSAM and people think is CSAM, and when held up next to the original verified horrific image everyone agrees is the same image. At this point, someone is going to ask, rightly so, where that came from. In order to generate this attack, you have had to go out and procure, deliberately, known CSAM. Ignoring that it would be easier just to send that to the target, rather than hiring talent to recreate the pose of a specific piece of child porn (or 30 pieces to trigger the reporting levels), the most likely person by orders of magnitude to be prosecuted in this scenario is the attacker.
- rlpb 5y ago
  > Upon visual inspection, your 'legal porn' is going to have to at least look passably like image 105 of the dataset to get anywhere. Define "get anywhere". Why won't you get raided by the police and have all your devices seized first?
  - shapefrog 5y ago
    No, probably wouldnt even get to the desk of the police. End of the road would be NCMEC, to whom apple would refer hash matching images that pass human verification that are close enough to porn. If your 30 or so hash matching images matched their corresponding known CSAM then that goes on to the police and then they knock on your door.
    - rlpb 5y ago
      > to whom apple would refer hash matching images that pass human verification I am under the impression that Apple's scheme allows them only to verify the output of the matching algorithm (the "safety vouchers"), and not the image content itself. So in the hypothetical situation described in the thread, it won't be possible for Apple to detect the false positive, and they could pass on the report to NCMEC. I fear that this will lead to a law enforcement raid without any actual human verification of the offending image itself. If I'm wrong, I welcome citations which demonstrate the opposite.

arn 5y ago

How would you know if it's a CSAM collision? I don't believe that database is publicly available anywhere, for obvious reasons.

arsome 5y ago
Apple's client clearly has some revealing information on that...
ALittleLight 5y ago
Good point. I can offer three possibilities to how you might know. First, a data leak of the material. Second, law enforcement presumably gives hashes of known CSAM to service operators so they can detect and report it. You could pose as or be the operator of such a service and get the hashes that way. Third, if you were a government operator you may have access to the hashes that way. (Although I guess corrupt government agents would have other easier ways of getting to you)
FabHK 5y ago
Yes. That, plus it's unclear whether you can create a collision with a nonce. It's a perceptual hash, not a cryptographic one. Lastly, to compute that collision might be so expensive that you could instead compute small SHA-256 hashes, ie mine BTC, and use the money to obtain your sinister goals via other means.

Ashanmaril 5y ago

If you're gonna go through that much effort, you might as well just text someone an actually illegal image from something other than an iPhone You're already attempting to frame someone for a crime, might as well commit another crime while you're at it

robertoandred 5y ago

Why would you save to your personal photo library some porn that a creepy rando sent to you in a text?

shapefrog 5y ago

Your hash will match to say image 105 of the dataset. Upon visual inspection, your 'legal porn' is going to have to at least look passably like image 105 of the dataset to get anywhere. So at this point we have an image that computers think is CSAM and people think is CSAM, and when held up next to the original verified horrific image everyone agrees is the same image. At this point, someone is going to ask, rightly so, where that came from. In order to generate this attack, you have had to go out and procure, deliberately, known CSAM. Ignoring that it would be easier just to send that to the target, rather than hiring talent to recreate the pose of a specific piece of child porn (or 30 pieces to trigger the reporting levels), the most likely person by orders of magnitude to be prosecuted in this scenario is the attacker.

rlpb 5y ago
> Upon visual inspection, your 'legal porn' is going to have to at least look passably like image 105 of the dataset to get anywhere. Define "get anywhere". Why won't you get raided by the police and have all your devices seized first?
- shapefrog 5y ago
  No, probably wouldnt even get to the desk of the police. End of the road would be NCMEC, to whom apple would refer hash matching images that pass human verification that are close enough to porn. If your 30 or so hash matching images matched their corresponding known CSAM then that goes on to the police and then they knock on your door.
  - rlpb 5y ago
    > to whom apple would refer hash matching images that pass human verification I am under the impression that Apple's scheme allows them only to verify the output of the matching algorithm (the "safety vouchers"), and not the image content itself. So in the hypothetical situation described in the thread, it won't be possible for Apple to detect the false positive, and they could pass on the report to NCMEC. I fear that this will lead to a law enforcement raid without any actual human verification of the offending image itself. If I'm wrong, I welcome citations which demonstrate the opposite.