This directly contradicts Tim Cook's statement, in fact he used the same term "master key":
> In the physical world, it would be the equivalent of a master key, capable of opening hundreds of millions of locks — from restaurants and banks to stores and homes.
It either is a master key or it isn't.
the FBI asked for a specialised OS image that would have a check so that it could only work on this specific phone. Combined with Apple's encryption/signing of the update, this image could not be applied to any other device.
So the "work" asked of Apple would produce something that would only work on the phone with the warrant
"Specifically, the FBI wants us to make a new version of the iPhone operating system, circumventing several important security features, and install it on an iPhone recovered during the investigation. In the wrong hands, this software — which does not exist today — would have the potential to unlock any iPhone in someone’s physical possession. The FBI may use different words to describe this tool, but make no mistake: Building a version of iOS that bypasses security in this way would undeniably create a backdoor. And while the government may argue that its use would be limited to this case, there is no way to guarantee such control."
The software asked for by the FBI would only work on one phone because it would both have a check for the specific phone, and be signed by Apple (so can't be modified).
The software would need to be modified and resigned by Apple to be used on another phone.
You really, honestly, truly and genuinely believe that if Apple just does this for them, just this one time, just for this one phone, then that's the end of it? That the FBI and other law-enforcement agencies would never, not once, not for the rest of the lifetime of the universe (and certainly not, say, about 24 hours later) come back and say "Well, we're glad you helped us with that case, now here's this list of other cases we'd like help with, and now you've proven that you can do this we're going to be bringing you a lot of these"?
That's what Tim Cook is worried about: it won't stop with this one phone unless it's stopped before that one phone gets unlocked and becomes the precedent that makes this just another routine thing. Every now-routine violation of privacy and security was originally marketed as "just this one case, this one's unique and special and sensitive and we absolutely have to get it done, just this one, trust us" and eventually morphed into "eh, it's Tuesday, time to ship off another truckload of requests to have this done to people".
The slippery slope argument being applied here is that FBI search warrants will now work on phones less than the iPhone 5S from now on (until Apple closes the iOS update backdoor I guess). This is far from "China will be able to mass-hack into iPhones from now on"-style comments I've seen from others.
The fact is that Apple left a backdoor in their phones and are now being told to exploit it. An argument against having the backdoor in the first place.
Furthermore, the US legal system encourages elaboration on stare decisis, applying prior case law to novel cases rather than hashing out new decisions. This means that creative applications of this ruling ("give me an uncontrolled backdoor") are simply a question of time once the landmark case is made.
What this case is doing is setting precedent that the courts can compel a company to _create_, no matter how trivial one may think that creation may be today. _That_ is the precedent so many draw exception to, because stare decisis is also a doctrine of incrementalism, of gradual expansion of interpretations. Today Apple is compelled to create a very controlled firmware; who's to say in the figurative tomorrow that Samsung won't be compelled to create and send a firmware update to a specific Blu-Ray player that creates an air microphone out of the laser? Where does it stop?
To be complete (as was pointed out in another sibling thread) incrementalism may be curtailed if implications are carefully argued and acknowledged by the judge. I am skeptical of the value of that approach, but have no hard argument against it.
You can't even downgrade to an old iOS version, because Apple won't sign them anymore.
At this point, the only valid answer from Apple is that any future devices (if they aren't this way already) should either require the password to do a software update, or wipe the phone.
The core thing that makes this not a master key is that Apple has to make the decision to comply on each phone (and warrant). Which gives them the opportunity to push back if it chooses to. If the FBI has somebody's phone and secretly wants to unlock it, it can't without going through the courts and Apple.
At that point we should expect more world governments to get the key. Where does that end? Maybe with more companies having to demonstrate that they defend against their own subversion.
Apple could sign it offline when installing to the device.
The signature is customised to each device it's installed on, and has a nonce so it can't be saved.
The US legal system and education pertaining to it is largely composed of studying prior case law and its applications. Few cases (including this one) involve truly new decisions, and even those are usually novel applications of former decisions. That is what those who can see past their nose are concerned about. There is every probability that, in the near future, one or both of Apple or the judicial system will tire of the farce of one-phone-per-case firmware and request a new application of the law (because there would now be precedent).
To take any laywer-moderated statement (including Apple's) at face value belies credulity that does not become the HN audience. As others have pointed out, Comey's appeals to emotion and terrorism should be enough, but the idea that an FBI lawyer would advise pursuing this particular case in this particular manner without expecting to set precedent? Positively silly. This is what they do.
If a new case is not similar to this case, then this case doesn't count as precedent. If a new case is similar to this case, then the arguments put forward for this case apply to the new case, and we shouldn't care about precedent setting. If the new case is similar but different in important ways, then when that new case is fought, the lawyers point ou the substantive differences, and it doesn't count as precedent.
The only concern would seem to be if the future court fails to analyse a case properly. But they can fail regardless of what precedent is or is not set.
I don't think individual cases should be ruled improperly because of concern about the precedent effect. You might want meta-level rules (e.g. Bill of Rights, free speech), but the individual actors should not be taking them into account except insofar as they have been legislated (although the Supreme Court perhaps should).
TL;DR argue the damn object level.
Whatever you think they should do, judges absolutely consider the precedent effect when ruling on accepting arguments, and lawyers most certainly consider precedent and setting it when presenting those arguments. There may be some intentional offloading of those decisions from lower courts to the appeals process, but no judge makes their decision in a vacuum. Nor should they.
I mention that only when people assert incorrect facts about the case. Correcting someone when they're wrong about a matter of fact does not imply any legal opinions.
About your point: there's a difference between "If I agree with argument X in this case, I should agree with it in that case; however, X shouldn't apply in that case, and I can make no principled distinction between the two. Therefore, X can't apply even in this case" and "I agree with argument X in this case, but not in that case; here's a principled distinction between the two; however, since it might lead to someone misinterpreting it down the line, I'm going to rule against X."
The first is a concern for precedent that I'm fine with, the second is not. My understanding is that judges will lay out the reason X applies here but wouldn't apply in the other hypothetical, which makes the concern about it being used as precedent later unwarranted.
The form of the argument here seems to be the second. "Yeah, this one is mostly fine, but we're fighting because of precedent. If they can force us to sign software and install it on a single phone, who's to stop them from forcing us to put software on every device we sell?"
And the clear distinction between the two is that one is only being put on phones that have a warrant (and belong to the government, to boot), and the other would also harm innocents. So the proper response should be "comply with this order, but specifically because of the fact that no innocents are being caught up", and that way it doesn't set any harmful precedents.
Or if you think there's a different reason why this case is fine but the general case isn't, then that itself is a reason for the general case not to have this case as precedent. Whatever those reasons are, make them explicit. If Apple puts those arguments into the court record, and the judge explicitly says "it's ok for X but not for Y",that defeats the harmful precedent.
Am I misunderstanding anything, or do we just disagree?
I think you're applying some fairly strict high-mindedness to the US legal system while others (myself included) worry that creative extension of intent appears rampant in cases that touch on technology or terrorism and therefore fall into your second form. Many of us don't, as a general rule, trust the courts, law enforcement, or our government to do anything but what is politically expedient and beneficial to them at the moment.
While the wording of this ruling is about one phone and a particular method the FBI has laid out, it appears to set precedent that law enforcement can compel a product company create a non-existent product (no matter how trivial) in order to exploit a known security vulnerability in one of their products. This is what concerns me and many others, because it brings us very short steps away from "make us a version that works against the Secure Enclave" to "make it work as an OTA update over WiFi" to then "make it work as an OTA update over cellular" and subsequently "make a version we can incorporate into a StingRay" and forward. None of these would be illogical steps to take in abject pursuit of stamping out terrorism and might even applauded by parts of society, but taken as what some perceive as an inevitable whole they paint a dim picture for personal privacy.
It's an improbable coincidence that the FBI has elected this particular charged case in which to stake their flag. Given their pleas with technology companies for cooperation over bypassing cryptography in recent months, this appears to be a logical continuation of that campaign.
In conclusion, we likely disagree, specifically due to my cynicism and your seeming lack thereof.
My main point related to this was above:
>The only concern would seem to be if the future court fails to analyse a case properly. But they can fail regardless of what precedent is or is not set.
To argue against that, you'd need to claim that precedent makes it easier for the later court to fail. Do you have examples, where it should have been clear that precedent didn't apply, yet the court reached the conclusion that it did, incorrectly?
(Preferably in important cases.)
>While the wording of this ruling is about one phone and a particular method the FBI has laid out, it appears to set precedent that law enforcement can compel a product company create a non-existent product (no matter how trivial) in order to exploit a known security vulnerability in one of their products.
I've said elsewhere that this argument seems to be useless. If Apple says it's "unreasonable" to expect them to do this, then they might be forced to hand over the source code, and the FBI will create it themself. The problem is
1. iOS is closed source and
2. iPhone requires a signature from Apple
If Apple doesn't help them, they could conceivably be forced to simply hand the keys and code over. It's a benefit to Apple to be able to create it themselves and maintain control over the keys.
>This is what concerns me and many others, because it brings us very short steps away from "make us a version that works against the Secure Enclave" to "make it work as an OTA update over WiFi" to then "make it work as an OTA update over cellular"
All of these seem fine, assuming that Apple is not modifying the phones to make it easier to hack. In other words, I agree that they follow as direct precedent from this case. (Although they can decide to only update a phone after a proper warrant.)
>make a version we can incorporate into a StingRay
This is the part that doesn't follow from precedent. Giving over control of the tool to the FBI who could use it without a warrant is novel, and would require a judge to justify it.
>In conclusion, we likely disagree, specifically due to my cynicism and your seeming lack thereof.
Maybe I'm just more cynical than you. You worry about a future court doing the wrong thing because they're misled by precedent here, I'm worried about a future court doing the wrong thing just because. I don't think the risk goes up significantly based on this precedent, because I think it could happen anyway.
>It's an improbable coincidence that the FBI has elected this particular charged case in which to stake their flag. Given their pleas with technology companies for cooperation over bypassing cryptography in recent months, this appears to be a logical continuation of that campaign.
Pure tinfoil material here. Do you know of any cases where the FBI had a phone but didn't try to unlock it, that could be said to be as urgent as this? This isn't the only court case with Apple going on, it's merely the one that got a lot of attention, and certainly much of that attention is Apple's fault (not all, but a lot).
No, I'm arguing that incrementalism means the court doesn't have to fail (as a court) - the follow-on steps that I outlined and you found perfectly acceptable will not stop wherever you personally think they will, nor have I outlined (or even imagined) all the incrementally creative applications of this case that will likely follow. I'm also arguing (which you didn't counter) that this case lays the groundwork for the FBI to request _new_ products from companies.
> This is the part that doesn't follow from precedent. Giving over control of the tool to the FBI who could use it without a warrant is novel, and would require a judge to justify it.
I was being brief to illustrate, not to set a literal expectation of progression; I'm not here to present an argument for court. Giving the FBI control over the tool could be argued at any point in the very long lifetime of this decision, be it tomorrow or years from now. Perhaps after the courts have decided to allow for multiple devices or when suspect X is holed up for weeks in their cabin eating pizza and playing candy crush and the FBI cannot reach them by normal investigative means.
> Pure tinfoil material here.
Perhaps. Consider the probable value of the data on the phone beyond what's already available externally. Consider that the FBI could have simply requested Apple provide the data on the phone instead of specifying the mechanism and thereby creating precedent that they can compel a company to create something new. Consider that, in spite of generally dim views of the FBI's technical capabilities, the FBI usually provides sound, conservative advice with regard to electronic evidence (as a forensic analyst I've been party to a lot). Consider that over the past year+ the FBI has been protesting with little effect in congressional hearings that $todays_crime_buzzword are "going dark", that technology companies need to cooperate with them, that "securely insecure" (my words) systems are possible if created "at the design level" (not my words). They were nearly shouted down in those hearings. Consider the subsequent public visit the FBI and others made to Silicon Valley, extending an olive branch and returning effectively empty-handed.
Then consider that this case, involving an older and still-vulnerable version of a technology the FBI has been warning about, suddenly falls into the FBI's lap. It's politically and emotionally charged, it involves terrorists - they can ask for the world and society will rubberstamp it "because terrorists."
So the FBI asks for something small but very specific, and in doing so "happens" to set a precedent that they can compel companies to create a new product to bypass the very security measures they've been arguing vehemently against. Do you really think that's accidental?
> I'm also arguing (which you didn't counter) that this case lays the groundwork for the FBI to request _new_ products from companies.
I don't really have an opinion on that. I haven't commented much on whether they should have that power.
I think in this case, Apple would prefer to make it themself than be forced to hand over the code and keys to let the FBI make it. If Apple says "this is too hard for us", it's plausible that they'll need to do that instead. I have seen this point made elsewhere, but I'm not sure of its validity.
>Consider that the FBI could have simply requested Apple provide the data on the phone instead of specifying the mechanism and thereby creating precedent that they can compel a company to create something new.
My understanding was that the FBI asked for the data, Apple said "we can't do it", FBI countered with "Do X, Y, and Z, it is technically possible". Is my understanding incorrect? It sounded like the order needed to specify exactly what was to be done.
About the precedent point:
I think the proper response is to argue against specific details of the actual case (which you have with the point about forcing them to make a product). If you don't have specific objections against the actual case, but worry about precedent, then make explicit the difference between the specific case and the general one, and try to get that acknowledged by the judge. Introduce those arguments in briefs to the court, and try to make it very clear to a future court what the limits of the precedent set should be.
If you haven't imagined what could go wrong, then don't reject something because of unknown dangers. It seems to be born of a lack of trust in future courts, but there's no particular reason to distrust future courts over current ones.
(If you happen to know any philosophy of law articles that discuss this topic, I'd love to see them.)
I don't think it's too hard for Apple to do this; they could probably turn out the request in very short order. However, as soon as they let this go and do that very thing, their legal obligations skyrocket as thousands of cases pour in; in truth, my cynical conjecture is that this is the very reason Apple is fighting this fight. Not for the consumer, but to minimize the very real cost associated with satisfying this kind of request. The Secure Enclave is likely as much a legal defense for Apple as it is a technological one for the consumer. Whatever the motivation, in this case the consumer seems to win.
I must admit I don't disagree with the specific details of the actual case. What I draw exception to is the specificity of the ruling (dictating how Apple does business) and the precedent it sets of enabling courts to force a company to create. That exception is exacerbated by this case's proximity to the FBI's very recent and very real behavior regarding cryptographic systems.
Edit: LOL, downvotes because you don't like it. Apple made a bad design decision, and it's coming to bite them in the ass. Yeah, it sucks, but I'm not the one who designed the backdoor. Blame the corporate culture, not the messenger. Fix the problem and tell your customers about it. Don't pretend it doesn't exist.
User confirmation has come in later OS updates it seems, so the actual backdoor has been closed.
The right thing would have been for Apple to unlock the phone in question, and issue a security advisory for the longstanding but newly-relevant vulnerability. Then, if they want their devices to be secure against USG, actually fix their security model to remove their own privilege. Instead, they walked right into the FBI's trap on a completely unsympathetic case. They're setting us up for a horrible precedent simply to continue pretending they can provide a level of security that they didn't actually build.
1) Identify target phone and obtain warrant to search.
2) Invoke Apple<-->FBI-specific api to initiate OS image generation to the specified target phone. OS image would be automatically downloaded to target phone.
3) Once downloaded, OS image would present FBI-specific API to allow exfiltration of any desired data.
This might not violate the 4th Amendment ("The right of the people to be secure in their persons, houses, papers, and effects ...") because they'll have a warrant, but IMHO it's a clear violation of the 5th ("... nor shall be compelled in any criminal case to be a witness against himself ...").
Yes, the particular OS image they produce would only work on that phone. But all of the work they did to enable the hack would work on every iPhone in the world, and the work necessary to take this OS image and use it on another device is effectively zero. As such, you can be absolutely certain that the FBI would start demanding Apple do this for other phones. And it's not just the FBI, every repressive regime in the world would start demanding this same capability. And Apple would have no grounds to deny it at that point.
"Specifically, the FBI wants us to make a new version of the iPhone operating system, circumventing several important security features, and install it on an iPhone recovered during the investigation. In the wrong hands, this software — which does not exist today — would have the potential to unlock any iPhone in someone’s physical possession.
The FBI may use different words to describe this tool, but make no mistake: Building a version of iOS that bypasses security in this way would undeniably create a backdoor. And while the government may argue that its use would be limited to this case, there is no way to guarantee such control."