4 ms·

N-gate summed it up very well: "if a guy in Cupertino can put a U2 album on your phone, he can just as easily put CSAM on it". If you are using a device manage

by piaste 3y ago

N-gate summed it up very well: "if a guy in Cupertino can put a U2 album on your phone, he can just as easily put CSAM on it". If you are using a device managed by Apple/Google/Microsoft, E2E is merely a roadbump - it means that your device owner needs to push a backdoored update to it, rather than simply reading your data server-side at their leisure. Which is still valuable, to be clear: it's a pretty large roadbump. Either you're such an important target that they're shipping a specific update to you alone (in which case, frankly, you didn't stand a chance), or they risk exposure. But if you install a little snitch that reads your messages and reports you if they match some complicated, necessarily hashed pattern, then that roadbump is mostly gone. All the guy in Cupertino now needs to do is update the patterns to match what they want them to match, without anyone noticing.

gickhocq 3y ago
I think I see your argument, are you saying that because the scanning involves transforming the content into a hash and comparing to a set of hashes, the problem is that this is resistant to third-party analysis? Not because of the overall behaviour of the scanner or indeed any other software on the device, which can be reverse-engineered using standard techniques, but because a database of hashes is inherently opaque as to the content it will match with?
- piaste 3y ago
  Correct. Is that hashed fingerprint you just downloaded associated with the face of a known victim of abuse, or with that of a whistleblower? Impossible to tell. (The N-Gate scenario is a little different: someone who doesn't like you controls a random app that is allowed to write to your /Downloads or /Pictures folder. Puts compromising material in it, then lets the well-meaning OS scanner pick it up and report you. You spend the next X years wrangling with law enforcement, trying to convince them that you were "hacked", the same thing every criminal claims.) If the scanner program merely refused to transmit the potential match, it would be acceptable. No avenue for abuse (without shipping a malicious update, which is always possible for cloud-controlled devices anyway), and at worst an annoyance in the event of a false positive. But I very much doubt that the parties pushing for this legislation would settle for that.

by piaste 3y ago

gickhocq 3y ago

I think I see your argument, are you saying that because the scanning involves transforming the content into a hash and comparing to a set of hashes, the problem is that this is resistant to third-party analysis? Not because of the overall behaviour of the scanner or indeed any other software on the device, which can be reverse-engineered using standard techniques, but because a database of hashes is inherently opaque as to the content it will match with?

piaste 3y ago
Correct. Is that hashed fingerprint you just downloaded associated with the face of a known victim of abuse, or with that of a whistleblower? Impossible to tell. (The N-Gate scenario is a little different: someone who doesn't like you controls a random app that is allowed to write to your /Downloads or /Pictures folder. Puts compromising material in it, then lets the well-meaning OS scanner pick it up and report you. You spend the next X years wrangling with law enforcement, trying to convince them that you were "hacked", the same thing every criminal claims.) If the scanner program merely refused to transmit the potential match, it would be acceptable. No avenue for abuse (without shipping a malicious update, which is always possible for cloud-controlled devices anyway), and at worst an annoyance in the event of a false positive. But I very much doubt that the parties pushing for this legislation would settle for that.