The "trying to find small numbers of people who might do some terrorist act via SIGINT" is post-9/11. That's when surveillance became dangerous. Now, law enforcement has access to far too much take. Worse, there's a huge US anti-terrorist establishment that's orders of magnitude larger than the number of terrorists and doesn't really have enough to do. So they self-generate work.
There wasn't much wiretapping in the US before the advent of electronic central offices. Phone switches just weren't equipped for it. Somebody had to physically connect wires to listen in. (Yes, you could listen to a few lines remotely using the test gear. Telcos hated that, because it tied up the automatic line insulation test equipment, COs had only 2 to 4 sets of that, each set took up three racks, and each could only do one line at a time.) In Guliani's old book about taking down the New York mob, he writes about the government having to pay phone bills to New York Telephone for wiretap lines. Once they didn't pay the bill, and the billing system billed the party being tapped. That's where the pressure for CALEA and built-in wiretapping began.
Project MINARET was one of the first times the public was made aware of the NSA. It started in 1962 and it involved the wiretapping of US citizens like Martin Luther King Jr and Senator Frank Church.
https://www.theguardian.com/world/2013/aug/23/nsa-prism-cost...
1. Claiming such things are conspiracy theories, despite an abundance of evidence 2. Not giving a shit, even during the (very brief) period where it was covered (very briefly) in the news 3. Buying into all the thingly-veiled gov and media FUD about terrorism, paedophiles, organised criminals and communists 4. Not protesting en-masse 5. Not voting for a party that runs on privacy and anti-surveillance policies 6. Not starting such a party as in (2), since AFAIK it doesn't even exist...
It's like if apple added a 360° motorized camera in your home but said "don't worry, it's only pointing at the door!"
There is zero breach of privacy unless the result of the scan are sent without your consent. Just like for everything else that's on your phone, scanned or not.
This is “simply” Apple leveraging people’s phones to run perceptual hashes rather than paying for the servers to do it themselves. Which is why the outrage caught them off guard, from their perspective nothing changed governments have been spying like this since 2011.
Those conditions would undoubtedly be tied to social credit scores, required viewing of ads, required on-demand "checking in" with the government, and direct access to local phone data on demand to catch and prosecute political dissidents sharing critical information and memes.
Example: Australians are now required to install an app that has them taking a photo of themselves and sent to authorities to enforce COVID mandates like staying in one location, even if you're not sick:
https://www.theatlantic.com/ideas/archive/2021/09/pandemic-a...
Semi-related - social credit score system: https://thepostmillennial.com/watch-australian-government-pl...
There is no changes whatsoever in that regard. You never owned your iPhone, and you still don't, no more, no less than before.
The question of whether we can and should build client-side scanning has been a huge and controversial issue in our community since at least 2019 — when AG William Barr wrote an open letter to Facebook requesting these features. As a cryptographer who works on E2EE I’ve been reading opinion pieces on this, writing pieces and attending Stanford-hosted workshops since that time. Many people attended those workshops from child safety groups, policy think tanks and other tech firms. Apple was also invited and declined to attend. But to imagine they didn’t realize how controversial these systems would be — that defies belief. They would have to be living in a cave, and the Apple security leaders don’t live in a cave.
The Apple system was clearly designed to enable scanning on E2EE backup files. The fact that iCloud isn’t E2EE right now is amusing, but I think it’s safe to assume that they want to enable that feature but felt they needed this coverage. Apple doesn’t share their reasoning, but it’s fairly obvious they felt some strong pressure — largely instigated by government agencies and members of the Senate - to implement some kind of “voluntary compromise” here. This is the one they chose.
Where they were surprised, I think, is that there was so much broad negative reaction beyond the usual suspects like privacy groups. As you said, server-side scanning is already ubiquitous so what’s the big deal? My suspicion is that “server side scanning of your private files is ubiquitous” is something that is only actually well-known to tech experts. Most of the general public is ignorant about all this, and they only really get a chance to react to any of this technology when something big makes it national news: as Apple did this summer. (Putting it on your phone and making it apply to private backup photos rather than shared albums, as some cloud providers have in the past, also didn’t help.)
As to scanning on device, that’s not what is involved in the child porn system as the list it’s being compared to is kept secret. So inherently the system requires an upload of each image for comparison. Further, it also needs to actually upload an image for comparison not just a hash.
This is a very dishonest take. They already have a system which runs on your device and processes all photos in the photo library, tagging categories like faces and pets and food and so on. They could have added this new system into that one with much less engineering and design effort using a system which scans all your files. They didn't, they went to a lot of extra effort not to. They already have system services e.g. location awareness or device monitoring and diagnostics, they could have built this new system as something like which would run 24/7 and access everything and be a simpler design, and they didn't.
What they did was put it into the iCloud upload feature to be matched by additional code on the iCloud server side. Something which took them more time, more effort, more complexity, and gave the whole thing less power and less flexibility.
If this system is as dreadful as all the complaints say, it should be plenty of ways to show that without having to make stuff up. If describing it accurately isn't scary enough, maybe it's not as scary as y'all want people to think.
In large companies -like Apple-, there is tracking of conversions (see the ad -> buy the product), churn (a client who stopped coming/buying), interest/reach (who am I targeting with this ad). And that's it. There is no tracking of "people who would have bought an iPhone but chose not to SPECIFICALLY because of the new tracking feature we added. Thinking this kind of metric is relevant and measurable is a complete fantasy, and borderline silly.
Public discussion might influence product decisions. But it have to be hugely controversial to have any effect.
You can't "vote with your wallet". That's it. I have no idea why such belief made it's way into pop culture.
If there's anything that will make a business listen, it's money (or lack of it). When it affects the bottom line, they are forced to listen.
Sure, I don't affect them much by myself. But everyone commenting about Apple around my presence will find out why I am not sponsoring them.
There is also the crappy hardware design, not just the privacy issues (which negate the advertising spending they made the months before). Here is what Louis Rossmann, who repairs Apple devices, says: https://www.youtube.com/watch?v=AUaJ8pDlxi8
It's not that the concerns aren't justified. It's that they're far too complex to be communicated and grasped by the public as a whole, and they facilitate far too much by way of immediate profit and market advantage to the vendors who will more than happily ignore the warnings.
A boycott by the vanishingly small fraction of the market which does fully grasp the issue is no threat at all.
________________________________
Notes:
"Wallet voters" live under the illusion that their intent is known and taken into account by BigCo decision makers. This is NOT the case. They just don't show up in reports. So what do you think they are going to do about it ? Nothing.
If you want to influence product design, you can talk about it online and condemn their choice. If that's big and controversial enough, they might do something about it. And ypu can do that while buying a new iPhone as well because it doesn't matter.
Isn't that how any vote works? My ballot is one of several hundred million - statistically meaningless by itself. But it is a vote. And if enough people vote, the millions of statistically meaningless datapoints become meaningful.
And in the end, even if my single wallet-vote is meaningless, I'm no longer affected by what Apple chooses to do with their phones. A successful vote, IMO.
Apple has been up front about what they're doing. They're explaining how what they're doing is specifically designed to protect your privacy in practical terms. No, they probably don't have to be doing this, and I would be happy if they decided not to.
But do you really think that any other phone you buy—apart from something ridiculously niche like a Librem—doesn't have exactly the same problems, only they're still keeping it secret from you?
"Voting with your wallet" only works if there are realistic alternatives.
At the moment, ALL of the other smartphone models (except iPhone soon) don't have this problem. There's no reason to bother with client-side given the current server-side implementations and this scanning can't be hidden for long.
https://www.oxfordhandbooks.com/view/10.1093/oxfordhb/978019...