It's reasonable because police forces simply don't have enough resources to abuse this system. As a result, innocent people are generally left alone and don't live in constant anxiety that their homes could be randomly searched by armed operators. That's exactly how things should work.

Compare that to remote backdoor access to everyone's encrypted communications. All they have to do is push a button in order to instantly get access to someone's mail, calls, instant messages, social media, phone location history, etc. Abusing this power is a trivial matter and there is little if any oversight. The Five Eyes intercept, decrypt and store even their own citizen's communications as a matter of course and will tip off law enforcement so they can parallel construct a case using means of investigation that are actually legal. Intelligence employees have been known to abuse their powers to spy on their significant others, a practice that became known as LOVEINT. People who are a threat to those in power -- political opposition, whistle blowers, journalists, dissidents -- will probably be targeted despite their rights to privacy.

In order to obtain encrypted communications, police forces should have to get a warrant and obtain the physical storage media that contains the messages they want.

>It's reasonable because police forces simply don't have enough resources to abuse this system.

Could this principle be applied to encryption? A cryptosystem for personal communications carefully designed to be crackable using certain amount of resources. So getting access to one person's data would cost you say $1000 of cloud resources. And first you have to obtain an escrowed hash to crack using a warrant.

Would this be acceptable to public?

It’s not only the factor that police raids cost money; raids are also very visible, and there aren’t enough police to conduct very many raids all the time, so there’s an inherent limiting factor. Both these factors make it impractical to conduct blanket raids on everyone all the time.

> Compare that to remote backdoor access to everyone's encrypted communications. All they have to do is push a button in order to instantly get access to someone's mail, calls, instant messages, social media, phone location history, etc. Abusing this power is a trivial matter and there is little if any oversight.

Much as you can build a physical policing system with or without oversight and with or without covert powers, you can do the same for encryption. It is absolutely possible to build a system where the company, the police, and the courts all need to agree in order to hand over access — in fact it's more possible than with physical security. It's also possible to build regulatory bodies that make relevant details publically visible.

Look no further than the hacking of Bezos' phone.

You might not be aware that there’s actually no solid evidence that Bezos’s phone was ever actually hacked.

Analogies make shit arguments when we can discuss the matter at hand directly instead. Systems with central access controls are vulnerable to at least 2 separate sorts of attacks. One procedural and one technical end.

On the procedural side its trivial to degenerate from

* subjects privacy is breached when the lawfully constituted authorities present proof to a judge of reasonable suspicion that subject has engaged in antisocial and immoral acts against society or its members

to

* it would be be awfully nice if we had access to everyone's personal data so that we could punish anyone we like for anything we like

In the United States we have already standardized on breaching everyone's privacy to the greatest degree possible with little or no recourse for the citizenry. Once you have all the information on whom speaks or believes in a certain way acting against them is a relatively smaller step.

On the technical side a centralized system is impossible to secure. When not if it is compromised to the victors go the spoils. It is comparatively harder at least to attack millions of clients even if in theory a central node could be used to compromise clients this can in practice trivially be made harder to exploit by just not allowing server to push to clients and not updating every second. Attacks may need to remain undetected for a substantial length of time before a substantial portion of the user base is poisoned and the chance of detection climbs towards 100% the more parties you attack.

Furthermore if you become 100% effective at detecting illegal porn shared via whatsapp doesn't it follow that users will use a decentralized means of transmitting illegal information. We would be substantially disadvantaging the population as a whole for little gain.

“Illegal porn” isn’t the preferred phrase in this case. It’s “child sexual abuse images.” We’re not talking about pornography here.

We ARE talking about porn I'm not sure how you can be confused on that point. Nobody is defending the act of abusing kids or scumbags sharing media of same. Both acts are evil.

The biggest point is that backdooring everyone's communication will do exactly jack to prevent child abuse. It wont even make it much harder for scumbags to share it even more of them will use just use privacy preserving p2p applications. In fact if you fuck with everyone's sense of privacy you will be liable to push a much much larger portion of the population to use such tech.

The greater number of normal people who aren't drug dealers, crypto nerds, or pedos using stuff like TOR the harder it will be to pick out the weirdos. Meanwhile most of the harm to children wont be captured on digital media because it will keep happening when trusted adults abuse minors they ought to be protecting instead of harming.

Attacking the sexual abuse problem in America by backdooring communication platforms is about as effective as reducing deaths due to traffic accidents by hanging out in parking lots.

> Meanwhile most of the harm to children wont be captured on digital media because it will keep happening when trusted adults abuse minors they ought to be protecting instead of harming.

Can you explain what you’re getting at here? I don’t get how this fits into your argument.

The problem they are highlighting is child abuse. Child abuse is indeed a horrible thing. It's also a very hard thing to combat because it largely happens when a trusted adult abuses a child and mostly doesn't document it for posterity or share it on the internet. The victims face the double whammy of shame and social pressure not to accuse the victimizer.

Attacking the perverts sharing pictures is worthy but does almost nothing to prevent child abuse. The best case scenario is you expose a small portion of perverts and thus prevent idiots from trusting their kids with them.

Unfortunately attacking messaging services will do little to combat perverts sharing pictures as they can trivially switch to using slightly better technology.

The effect of multiplicative. Imagine you start with 1% of victimizers sharing media online. Imagine that you catch 1/10 of 1% of them before the rest figure out centralized sharing is not the best idea. You are now stuck with permanent downsides and will catch few additional people.

0.01 * 0.001 = 0.00001

The logical conclusion is that this isn't a very great way to combat child abuse.

Great! But you seem to be the one who’s confused. I was commenting on semantics. The “illegal porn” being referenced above is not porn. It is abuse.

Carry on with your discussion of why backdooring encryption is bad. I wasn’t commenting on that matter. This entire thread has no real discussion of how to solve the “encryption-while-scanning problem.” With the massive growth in CSAM sharing in the last few years, enabled by easy to use services like facebook and dropbox, it’s clear that solving this dilemma is just as important as protecting some normal person’s ability to use TOR. I’m extremely privacy focused, but I’m curious to learn if crippling fb/dropbox/etc is worth it for the sake of solving this issue.

Also, “traffic accidents” isn’t preferred, either! Do you hate me now? Language is important. I’m a transportation planner who focuses on safety, so I can talk about this all day.

You CAN'T solve the encrypted while scanning situation. AI is capable of misidentifying a naked baby picture as porn. Nobody will want to run a piece of software that might accidentally report them to the FBI to have their life ruined. So long as users have the ability to write and distribute software from regions outside your jurisdiction and users have the ability to install software of their choosing you can't stop people from communicating information you don't approve of. This is equally true of political discourse you and I both agree ought to be freely disseminated as is of evil material you and I would both desire to ban.

Maximal crazy is implementing 1984 in order to pretend to stop child abuse while it continues to go on all around us.

> In the real world, the police can get into anywhere, and get basically any physical object, once they have a warrant.

Objects/documents aren't analogous to walls.

If you were required to have transparent walls then the police could put cameras everywhere in the street and use them to record everything that happens in your house. If your walls are opaque then they can't. Even if they get a warrant, there is no way to go back and look at a recording of what happened inside your house the day before, since the opaque walls prevented any such recording from being made. It's well analogous to transport encryption.

Consider this: If encryption is suddenly turned off for all communications, and police have access to all of it via warrants, that's like getting a warrant for every home in the US. Obviously if the police had a perpetual warrant to your home, you wouldn't want that, why is it reasonable they have one to your data?

And it has been proven many times that not only will the access be abused, the secret backdoor will not remain secret and now every unscrupulous person can potentially have the same access.

The physical world is tangible and can be accessed by pure force. The mind and thought, and by extension our intangible data, are not subject to the same rules and are (and hopefully will remain) a safe haven of privacy.

There will be no more freedom if we can't keep our thoughts secrets from the authority.

I feel like there is a difference between having a thought, and expressing that thought in (entirely tangible, in terms of electromagnetic signals) data over a wire that someone else owns. If you want your thoughts to stay yours, you’re free to keep them to yourself; just don’t transmit them through electromagnetic media that other people currently control by force.

To make this analogy work, the government is asking everyone to install new front door locks. They have a master key. They will always have a master key. They can stop by whenever they please, if they suspect a crime. If you run a red light, they can search your house (no such thing as a "limited" backdoor -- they can't partially decrypt a hard drive. They just clone the whole thing).

Except that it's more like: In addition to the locks that they can access anytime, they're also installing CCTV, microphones and other sensors so they can be sure you're fully compliant at all times.

I do not see it that way. In years past if one did something he thought government might find offensive he could hide the evidence. Which was perfectly fine -- one would not be jailed for refusing to provide self incrimination. "Go search, but I will hide it; I never had it by the way".

Perhaps a better analogy: we haven't seen legislation that attempts to outlaw physical safes that are capable of destroying the contents in response to an attempt to tamper with or force them open, have we?

Such safes are very uncommon. If some major bank announced that it's going to install destroy-on-tamper on all its safety deposit boxes, I'd definitely expect some reactive legislation about it.