I don't know why they are not proposing this instead (well, I do, because they'd prefer more power), but I do not find this proposal acceptable either.

The fact that you might be able to see that a government read a message of yours offers little assurance that the government will not start abusing this power. After all, what possible recourse do we have after we start seeing these notifications on our messages? Certainly the government won't start telling us the exact reason they are snooping. The very fact that your conversations had been snooped on by the government might imply there is something shady about you. Where there is smoke, there's also fire, right? And this still has us relying that governments won't start pressuring clients to have special code which bypasses the notification when it is convenient for them.

No. The government must not be allowed to deny the citizen his right to have private conversations.

> little assurance that the government will not start abusing this power.

My hope is that if millions of people started seeing these messages next to their conversations with their lovers, the people would vote for a different government.

There's a big difference between knowing big brother is scanning for criminals, to seeing big brother has right now looked at my private message.

That's a bleak future to hope for. We can do better. How about if we do not allow big brother to scan for criminals in the first place? Big brother has no business scanning for anything.

Also, remember that criminals are defined by the law and the law is defined by the government.

The key to all this is probable cause. Which is exactly what the mass data harvesting operations are seeking to circumvent.

But they already do that with warrants. As long as they have a warrant and aren't breaking encryption by putting in backdoors, they should have the ability to listen in on conversations.

That's how it is now without encryption.

Yes, without encryption, they have that possibility, after they prove probable cause. Without getting into a debate whether that is appropriate, it is important to keep in mind that it is the citizen's right to privacy of conversations is the more fundamental here. The government does not have a right to snoop on everyone's conversations, they simply have this possibility in certain situations.

So, for me, this changes nothing.

The obvious downside is that anyone who cares about privacy will write their own client, governments will regulate against unofficial clients, platforms will comply with the regulation and lock everything down, and the end result will be users unable to exercise any control over what's installed on their devices.

I can picture this. It's the year 2050. You hear about how a dangerous man was arrested the other day and is being charged with owning an unlawful communication device.

This is known as a key disclosure law. We have such a law already in the UK.

* https://en.wikipedia.org/wiki/Key_disclosure_law

* https://www.schneier.com/blog/archives/2007/10/uk_police_can...

* https://www.theregister.com/Print/2007/10/03/ripa-decryption...

> Why isn't this middle ground being discussed?

Why should it? What problem does it solve? Certainly not the problem of terror in Vienna.

I would especially like to see a cryptographic mechanism enforcing such middle ground. That is, a mechanism that would allow law enforcement accessing the private key or plain text, but only with some inescapable side effect which would hinder abusing that power. That may be producing a cryptographic "proof of compromise" for the person being spied on, or the spying being publicized, possibly with some delay, or something else. I am searching for an analogue of having police show up for a home search, which would notify both you and your neighbors, and cannot be done sneakily and by-the-book at the same time.

Are there some interesting candidates for such a mechanism? At first is sounds like a long shot, but there are cryptographic mechanisms achieving unintuitive results, so it may very well be possible.

But then again there is a mechanism already, just not cryptographic, and that is police having to physically seize your device. I guess we don't need to work on this. There's nothing special about an E2E communication device compared to other things you may have in your home.

It is much more difficult to get data off an E2E communication device these days because mobile OSes use cryptography to mitigate the consequences of loss and theft. That is why law enforcement agencies want "the support of service providers," who can do things like deploy a backdoored WhatsApp binary, or bruteforce passcodes without triggering data loss. https://en.wikipedia.org/wiki/FBI%E2%80%93Apple_encryption_d...

That system simply isn't possible, or at least there are no candidates for such a mechanism at this time.

> Why isn't this middle ground being discussed?

that's what they need to do even today after they have found enough evidence that the court sent you what's called "notice of inquiry" and preliminary investigations start (it doesn't mean yo are guilty, it means there's gonna be an official investigation on you and after the investigation has ended they're gonna either drop the charges or go to trial).

but crime prevention usually works best when investigations are kept secret.

Imagine if in Italy we called every mob boss the police was investigating on to tell them they were tapping their phones or checking their bank accounts.

They knew it anyway, but that's where the middle ground is.

Another possibility would be, as I've written somewhere else, that a pair of keys is always created and made available to the local authorities, but they have to be authorized from a judge to use them.

The encrypted data is instead kept on the provider side, so that even if someone steals the keys, they can't access the data because it is in a separate facility, protected in the same way it is today, because it's in the best interest of the provider, that is already collecting that data, to keep it safe.

And if someone steals the data, can't decrypt it.

Metadata are also very important to investigators, I imagine that their need to decrypt the conversations comes from the fact that with the metadata they can narrow down the number of suspects to a few, but without the actual content they can't tell exactly what's going on.

> if they read the messages and find you've done some crime, authorities will eventually track you down

that's not so obvious.

I could show you a lot of cases when this has not happened, for various reasons, most of them formal (tapping lasted a few hours more than authorized, tapping was badly misreported in some other meaningless section of the conversation but it's enough to nullify it in its entirety, tapping included someone else that was not under investigation that should have been removed from the records bu it's still there, etc. etc.)

Can't tell if you are serious. What is the point of spying then?

Reminds me of that law that says if you plan to spy on the US you have to register on a list first. https://en.wikipedia.org/wiki/Foreign_Agents_Registration_Ac...

This is a good middle ground for espionage.

Even in the case of an investigation, there's the option for the authorities to request a time lag.

There's lot of options, we need to discuss this in depth weighing them all.

I think this sounds like a fair proposal