shuckles

61 ms·

901.
by shuckles 5y ago
You argued that somehow Apple blindly followed “elite consensus” to end up with the system they announced. I pointed out that the system they announced questions every aspect of “elite consensus” about how to scan for CSAM in a cloud servic
902.
by shuckles 5y ago
In addition, you can have expectations of privacy from far more than state actors that can coerce multinationals into violating unambiguous statements about where they draw a line in the sand.
903.
by shuckles 5y ago
This is a bad take, given that questioning the “elite consensus” is most likely how Apple arrived at this solution. Their proposed system doesn’t take for granted a whole host of assumptions made by every incumbent CSAM detection framework.
904.
by shuckles 5y ago
What use is all that goodwill if a solution which attempts to thread some incredibly complex trade offs in a way that innovates on the status quo in every possible way is met with skepticism that is rooted in mistrust of government and not
905.
by shuckles 5y ago
For ideological consistency, are you dumping every service provider that scans the contents of your account and reports offending data to law enforcement?
906.
by shuckles 5y ago
There wasn’t such a hurdle before or in the counterfactual where they built infrastructure to scan iCloud while also keeping iCloud Backups for every device.
907.
by shuckles 5y ago
That is literally the status quo with every cloud service. Apple, unlike the others, has said that they will evaluate you on the basis of what’s included in the associated data of your safety voucher, and you can inspect those contents beca
908.
by shuckles 5y ago
I’m not sure how you can conclude the speculation is “right” without engaging with the fact that this hypothetical is addressed directly in the threat model document and hasn’t been pulled off successfully against any of the other services
909.
by shuckles 5y ago
It’s not security through obscurity. The implementation is not necessarily a secret; it’s the model configuration which can only be differentiated if known. Since part of the threat is adversarial embeddings, it’s perfectly fine to assume y
910.
by shuckles 5y ago
In some ways, you can start to see the value in Apple’s system which lets the device user inspect what is stored in the associated data for later review.
911.
by shuckles 5y ago
Those services will take your account through the same, if not more invasive, process if you are found with a hash match like the ones being proposed in these comments. Unlike Apple, they’ve built interfaces that surface all your account ac
912.
by shuckles 5y ago
Correct, gradient descent attacks like this depend on being able to differentiate your hash function.
913.
by shuckles 5y ago
People here are proposing intentionally creating image assets which collide with perceptual hashes of known CSAM (ignoring whether that is legal or ethical) and sharing those assets to effectively SWAT unaware targets.
914.
by shuckles 5y ago
Counter-point: hijinks like this are defeated by including the original image instead of the image derivative in associated data. At that point, the system works in the exact same way as the photo scanning status quo.
915.
by shuckles 5y ago
> First, as an additional safeguard, the visual derivatives themselves are matched to the known CSAM database by a second, independent perceptual hash. This independent hash is chosen to reject the unlikely possibility that the match thr
916.
by shuckles 5y ago
Those will trip up 2020’s systems as well!
917.
by shuckles 5y ago
Why wait? Just send them the pictures on Facebook Messenger or Gmail or Dropbox today.
918.
by shuckles 5y ago
I’ll take the other side of this and bet that Apple offers the ability to remove themselves from escrowing iCloud keys before they expand CSAM detection to other kinds of content. Nerd points are at stake.
919.
by shuckles 5y ago
Presumably the system is novel enough that a very similar proposal which lacked one important insight (using perceptual hashes to turn image similarity problems into set intersection problems) was accepted to USENIX this year: https:/
920.
by shuckles 5y ago
This is quite off-topic from the fourth amendment workaround originally insinuated. I agree modern society depends on trust more than one may expect or wish. DDoSing the USPS to prove it doesn’t seem like a very pro-social idea, and I do li
921.
by shuckles 5y ago
This version is incorrect because Apple’s solution by design does not reveal the non-suspect contents of your luggage.
922.
by shuckles 5y ago
Known CSAM detection was one of three features they announced under the banner of child safety. Certainly they understand it is by itself an incomplete intervention for a much larger problem.
923.
by shuckles 5y ago
Is there a consumer computing device that’s more secure than iPhone against state actors? A Chromebook, maybe?
924.
by shuckles 5y ago
What abuses apply to CSAM scanning in this hybrid pipeline which don’t apply to iCloud Backup? If they scanned on the server, why couldn’t governments “slippery slope” abuse the system by requiring that all files on iOS end up in iCloud Bac
925.
by shuckles 5y ago
This is already the world we live in, with vendors like Thorn building SaaS solutions for people who accept UGC but don’t have the clout to integrate with NCMEC and Microsoft PhotoDNA directly.
926.
by shuckles 5y ago
I found it interesting that only ~1k of the ~500k banned words from WeChat were also censored in this engraving UI.
927.
by shuckles 5y ago
This looks like a good way to get a package marked as suspicious, and suspicious packages have their contents searched.
928.
by shuckles 5y ago
USPS x-rays many packages - I’d venture almost all of them shipped from a major population center.
929.
by shuckles 5y ago
CSAM in the database is confidential, so the human reviewer just needs to be plausibly confident that they have CSAM in front of them. However, it’s not clear to me that you can pull off all three simultaneously. Furthermore, the attack doe
930.
by shuckles 5y ago
I believe something like: an adversary creates an innocuous photo where: * The photo itself is benign. * The photo’s NeuralHash matches known CSAM. * The photo’s image derivative is not benign. It looks visually like CSAM. * The photo’s ima

More ›