Opinion: The secret gag orders must stop | Hacker News Reader

Opinion: The secret gag orders must stop

washingtonpost.com

204 points1cvmask48 comments

Good points. Secrecy was once used so that people wouldn't destroy the evidence the warrant sought. But now it's easy to tell the third party to retain the evidence; even if you delete it, they can still keep a copy. So there is no reason not to tell the target; they can't change the outcome, but they can begin the counter legal process to ensure their rights are preserved.

Sometimes I wonder if they want secrecy so that criminals don't know they're being investigated, and commit a crime that's easier to prosecute. If they just stop committing crimes, then there's no fancy press release saying how great the DA is or whatever.

An alternative of that is as Oppenheimer once said, sometimes things are secret because a man doesn't like to know what he's up to if he can avoid it

> If they just stop committing crimes, then there's no fancy press release saying how great the DA is or whatever.

That's an interesting observation on how people respond to their incentives.

If I'm a DA, my job description is probably not to reduce crime. It is to successfully prosecute crimes that have already been committed.

Of course, as a thoughtful and honorable person, I will naturally want to see if I can also reduce crime, not just prosecute it. But that is not the direct incentive I'm given.

As a software developer on various mid-size teams, I've often found that I don't get much of a boost in my performance reviews when I find a way to make the entire team more productive, e.g. by solving a nagging problem that is affecting everyone.

Instead, I get more highly rated by just making sure I knock out all of my own Jira tickets for the current sprint, and ignore those more global problems.

As a thoughtful and honourable person ... my guess is that there would be probably 6-12 months fighting the incentive then a process something like:

* Develop a philosophy that there are bad people out there.

* Greatest good strategy is to identify and incarcerate/re-educate these bad people through punishment.

* Therefore preemptive crime reduction measures don't really get to the root problem - they provide cover for the bad people to operate.

* Anyone talking about how all people are made up of shades of grey doesn't understand the situation and should be ignored. People providing legal support to the bad people are troublemakers.

* Bad people are the ones least able to resist prosecution. This is a mystery the DA never really sits down to rationalise.

That transformation isn't a new thing. Incentives and a smart mind can rationalise much harder situations.

> If I'm a DA, my job description is probably not to reduce crime. It is to successfully prosecute crimes that have already been committed.

Well, it gets worse. Look beyond the job description to the actual incentives people face.

If you are a DA with political ambitions, you might want to prosecute some very specific things, whether they are crimes or not.

(And also https://en.wikipedia.org/wiki/Parkinson%27s_law for empire building.)

> Sometimes I wonder if they want secrecy so that criminals don't know they're being investigated, and commit a crime that's easier to prosecute.

This is idiotic. When I worked murder investigations that used wiretaps, we didn't want the suspect to commit another murder. We wanted to gain enough evidence to see the suspect convicted. It defeats the purpose of the wiretap if they know we're listening to them.

Interesting relevation is the NY Times and Google gag order was actively contested. The contract between NY Times and Google required that Google disclose any data requests to allow NY Times to contest them. Google actively contested the gag order because of this clause. That is going to be requirement in any service provider contract that I sign going forward.

outstanding point

The 300,000 warrantless secret gag orders issued by the FBI [1] are why we should be E2EEing all the things. Really disappointed that Apple held off implementing further E2EE in iCloud at the FBI's behest [2].

A lot of very personal data is not E2EE in iCloud [3]:

- Calendars

- Contacts

- iCloud Drive

- Notes

- Photos

- Reminders

- Voice Memos

- Mail

This is honestly completely unacceptable. The current 'E2EE' is basically a farce - storing this data without E2EE in the cloud creates an extremely powerful state surveillance apparatus subject to gag orders. It is irrelevant that my Memoji are E2EE when my photos, notes, files, and calendars are not.

Furthermore, the E2EE implementation might not even be true E2EE:

> When a user enables iCloud Keychain for the first time, the device establishes a circle of trust and creates a syncing identity for itself.

> When the user turns on iCloud Keychain on another device, iCloud Keychain notices that the user has a previously established syncing circle in iCloud that it isn’t a member of. The device creates its syncing identity key pair, then creates an application ticket to request membership in the circle. [4]

It follows that Apple or a malicious actor can add an invisible device to the initial keyring server-side - a virtual 'zeroth device' - and 'approve' the user's very first device. It would receive everything added to iCloud Keychain. It isn't stated if the iCloud Devices list in Settings is a cryptographically verified list of all devices. Until Apple verifies that this is the case, we cannot reasonably call iCloud Keychain E2EE.

[1]: https://www.eff.org/issues/national-security-letters/faq

[2]: https://www.reuters.com/article/us-apple-fbi-icloud-exclusiv...

[3]: https://support.apple.com/en-us/HT202303

[4]: https://support.apple.com/guide/security/secure-keychain-syn...

+ my most sensitive service, Messages (if iCloud Messages is enabled)

Messages in iCloud is E2EE, but if you enable iCloud backup it's not:

> Messages in iCloud also uses end-to-end encryption. If you have iCloud Backup turned on, your backup includes a copy of the key protecting your Messages. [1]

I love it how doesn't state that practically speaking, iCloud Backup is not E2EE. It seems Apple is afraid to be forthright the privacy implications of iCloud Backup.

[1]: https://support.apple.com/en-us/HT202303

How do you expect a service used by average joes to exist without Apple holding a key somewhere? For 99.99% of the users of iCloud, losing access to their data because they lost the key is going to be a problem for them before having their data subpoenaed by a secret warrant is a problem.

Social/multi-key recovery ala ADSS? Or make it so as long as you have a single device you can perform recovery? It is very unlikely that you'll lose all of your devices.

That's half the point of the keychain syncing circle - even if you reset the iCloud account's password and kick every device out of the circle, it can then be re-established and you lose nothing.

And even if you don't have access to any of your devices anymore, as long as you remember just one of their passcodes (and have access to the account's recovery phone number), you can still recover the keychain.

> And even if you don't have access to any of your devices anymore, as long as you remember just one of their passcodes (and have access to the account's recovery phone number), you can still recover the keychain.

This implies a really easy brute-force! I doubt this is the case!?

You get 10 tries and then it's wiped from the cloud. [1]

Obviously this depends on you trusting that the HSMs function as Apple says they do and that it's their public key that you're additionally encrypting the keychain's key with. [2]

[1] https://support.apple.com/guide/security/escrow-security-for...

[2] https://support.apple.com/guide/security/secure-icloud-keych...

This is kind of terrible as it relies on an authentication barrier as opposed to a cryptographic one. No one can reasonably call this E2EE.

I really would have expected better security engineering from Apple, wow. Not that it matters, since barely anything of value is E2EE to begin with.

Thanks for the sources.

Like this: https://security.googleblog.com/2018/10/google-and-android-h...

TL;DR Android backups are end-to-end encrypted, but can still be restored to a new device without the involvement of the original device, using only the user's existing screen lock code and nothing else. Brute force attacks against the short screen lock code are prevented by secure elements in the datacenter.

Apple could have done this too but they caved when the FBI told them not to. They thus retain the ability to read all of your data including supposedly "end-to-end" encrypted iMessages, unless you disable iCloud backups completely. And of course Apple tolerates no alternative cloud backup apps in their closed ecosystem so you're SOL if you want end-to-end encrypted cloud backups. I think this is shameful for a company that makes such a big deal about supposedly caring about your privacy.

A setting called “use end-to-end encryption for all backup data” hidden in a submenu, with appropriate warnings all over it to make sure average Joes don’t turn it on. (Hell, make people buy a developer account if you’re really worried.) The fact that Apple won’t allow e2e even as an advanced feature is pretty telling.

I'm skeptical whether E2EE is the right lever here. We've seen that a sufficiently motivated government will figure out a way - exploiting security holes, backdooring personal devices, all the way to physically tapping fiber cables.

E2EE is the software engineer knee jerk technical solution, but you need to keep in mind you're dealing with an entity that has for all intents and purposes unlimited resources, long term motivation, the ability to force co-opt more or less everyone, and can amortize the cost over many cases. Which is to say: in the long run I'd bet on E2EE not working out.

And that's not to speak of all the technical disadvantages and complexity of E2EE, e.g. around searchability.

The right tool to solve this problem is (constitutional) law. We need to enshrine appropriate rights, and in a way that cannot be worked around with legal tricks.

Also, maybe controversially, there are cases where I believe it's entirely legit for the government to seize communications (some subset of severe criminal cases), once approved by a proper judge, with due process. I'd rather live in a society where there's a defined process for this, with appropriate legal scrutiny over both process and the targets, then in a society where we attempt to fight the government and it's a constant arms race. Not having a process means certain motivated actors get to cry "somebody think of the children" and then throw all legal protections over board.

I am speaking from the perspective that all governments are corrupt, and as history shows, all governments can and one day will betray the trust of their citizens.

If we want to preserve our privacy, then solid cryptography is our best bet. On the software side we are eliminating entire classes of vulnerabilities with advancements like memory-safe languages and strong containerization. The government would end up burning all their zero days if they attempted surveillance on anywhere near the dystopian scale of today.

I agree that these rights should be enshrined in law, but practical reality means that we have to take our privacy into our own hands.

> I am speaking from the perspective that all governments are corrupt, and as history shows, all governments can and one day will betray the trust of their citizens.

I think it's worth being carefully about not falling into absolutes. Yes, some governments, or rather some sub-parts of some governments, are corrupt. Yes, over centuries expect things to go south. But existence proof that some government somewhere sometime was corrupt does not mean all are, and doesn't mean you can never trust rule of the law.

> The government would end up burning all their zero days if they attempted surveillance [...]

The government will buy a $5 wrench, https://xkcd.com/538/. The only thing that keeps the government from doing that already is the rule of law.

Cryptography is a good protection from illegal actors, or from foreign powers etc; if you're fighting something that has immediate physical power over you, I believe the best bet is enforcing limits for the government in a democratic society.

The Fourth Amendment exists in US.

Find someone to enforce it.

Right, but it's not that black and white. There are exceptions for law enforcement, and details as to where it is applicable.

My point is that the solution is in politics and law making, to regulate these details.

Case in point: in my continental European home country, cloud emails are protected the same way as all your other communications, where in the US it appears they are not (?).

Which is to say: similar basic rules may yield pretty different outcomes, and the difference I believe is in politics.

The point is that in the US, there are not exceptions to our constitutional rights, but the government itself routinely engages in unconstitutional and illegal covert surveillance of citizens.

We already have the rights you discuss enshrined in our founding document, but the government just... ignores it.

At this point the only real solution is well-applied cryptography.

> the government itself routinely engages in unconstitutional and illegal covert surveillance of citizens.

This must be the subject of legal and political debate, right? Otherwise people would sue in the Supreme Court (or have already done that)?

> At this point the only real solution is well-applied cryptography.

See the other comment about the $5 wrench. I believe cryptography at best can attempt to patch some failures temporarily. The only long term fix is fixing the government.

"an entity that has for all intents and purposes unlimited resources"

When discussing cryptography, it makes a difference whether your adversary has "unlimited" resources, or just a lot of resources. There are challenges in cryptanalysis that cannot be solved even if you throw a planet-load of resources at them; but given literally unlimited resources, are simple.

"I'd bet on E2EE not working out"

You mean, you'll bet money that a government can eventually technically defeat any cryptography? Or do you mean that a government will do an end-run around E2EE?

If you mean the former, I'll take your bet; one beer-voucher says you're wrong. It's easier to devise strong ciphers than to crack them.

If you mean the latter, then I'm not playing. Political and social developments "in the long run" are too chaotic for me to want to wager.

> You mean, you'll bet money that a government can eventually technically defeat any cryptography? Or do you mean that a government will do an end-run around E2EE?

Hard crypto is hard, no debate there. My point is that the government will find a way around, e.g. by coercing you to give up keys, coercing an intermediate to backdoor your device(s), hacking your device etc. You're facing an APT that also has physical power of you. The only winning move is getting control over the APT/your government through the rule of law.

I've never been much of a fan of legislation that touches the internet, because it's invariably bad law, drafted by people who are afraid of the internet because they don't understand it.

You're right, cryptography isn't a solution to all privacy problems; you also need to conceal from your government/APT the very fact that you are communicating. OPSEC in general, really.

Concealing doesn't work if you communicate in public or semi-public settings (e.g. a WhatsApp group). I think that's a deal breaker for many if not most people.

The first Warrant Canary just turned 15 years old:

https://twitter.com/rsyncnet/status/1387090538273206274

It will be updated tomorrow morning, just like every Monday morning since April 2006.

Warrant canaries are unreliable. Apple, for example, stopped issuing warrant canaries in Sept. 2014 [1], but was added to the NSA's PRISM program almost 2 years earlier, in Oct. 2012 [2].

But I suppose a warrant canary by nature requires companies to be honest. There's not much you can do if the company is being deceptive with a vaguely phrased canary.

[1]: https://arstechnica.com/tech-policy/2014/09/no-apple-probabl...

[2]: https://arstechnica.com/tech-policy/2013/06/new-leak-feds-ca...

I have it on good authority that this one will continue...

What will you do if the government coerces you into falsely renewing your warrant canary?

I believe his defense against this is having three locations around the world that all need sign off for a canary to be published.

So there is no „the government“.

At least we would eventually learn that the government can do that, and fight it in court.

Secret gag orders are a dangerous game. It only requires one curmudgeon to bring it down. Some people legitimately don't care about future impact.

https://archive.is/9aZH3

Secret gag orders and secret courts shouldn't be a thing in countries that call themselves free. It's a violation of due process. For all I know, the worst conspiracy theories I can imagine could all be true just because secret gag orders exist. It's impossible to prove that they're not true when policy like that is acceptable.

It's an affront to democracy.

Illegal spying should remain illegal. When a spy agency has nothing to loose anymore, does not have to tread carefully anymore to see which laws to break or sidestep there is no bound to the eventual corruption that will come with time

We don't know the best arguments against secret gag orders because of secret gag orders.

Good luck with that.

Secrecy is a drug that governments become addicted to, and with zero down side for the government, nothing ever gets wound back.

Things will only get ever more secret.

We live in a hodge-podge of privacy violations.

The government abuses privacy. Private industry (especially Big Tech) violates privacy.

It will take a new generation of courageous politicians to sort it out. Until then, we're in for more of the same.

> When it comes to cybersecurity, the cloud bolsters protection. But now we’ve learned that the Trump Justice Department exploited this feature as part of a secret effort to obtain emails...

So he's saying that it doesn't bolster protection. And he's blaming the government for using "the cloud" in exactly the way that it was intended to be used: to spy on people.

I interpreted "the cloud bolsters protection" as cloud hosted solutions tend to be more secure since there is a single company which specializes in that service. They can protect the service better than any single company which develops it in house probably would since that is what they specialize in, it is managed, they have more money to throw at it due to many customers, and they can take advantage of economies of scale.

However that does make the cloud provided service much more of a target for those seeking that data.

Almost a courtesy that there was such an egregious absurd abuse of power. Can hardly ask for a more textbook example of how surveillance capitalism and authoritarian governments collide.

Reply on news.ycombinator.com