Proposals to ban working cryptography were all the rage in the Clinton years, but then they fell out of vogue for a decade, only to come roaring back in the form of bizarre proposals each stupider than the last, with Australia bringing home the gold in the Dumbfuck Olympics.
One feature of all this foolishness is the oft-repeated claim that it is possible to produce cryptography that fails every time the cops need it to, but never fails when criminals, spies, stalkers and identity thieves need it to. This is so implausible and obviously wrong that even people who only have the vaguest idea of how crypto works still immediately grasp that it is either a) bullshit or b) wishful thinking.
Enter William Barr, the new US Attorney General, who also wants to ban working cryptography. Though Barr's desire to ban working crypto is no less deadly and awful than, say, Rod Rosenstein's, Barr is doing one thing different: he's admitting that banning working crypto will make all the people who rely on crypto less secure, with the "trade off" that it will make society more secure because cops will find it easier to spy on "bad guys."
Even if Barr less wrong than other people on his side, he's still wrong. For one thing, it's impossible to keep working crypto out of the hands of bad guys, because working crypto tools are made all over the world and are licensed as free/open source software, and they run on any general-purpose computer, which is every computer, so as a practical matter, any ban on working crypto will only work if people who are willing to commit acts of terror and other bad deeds are intimidated at the thought of facing civil penalties for installing illegal software.
What's more, Barr is proposing that working crypto still be available for "nuclear launch codes" and other sensitive applications, including "customized encryption used by large business enterprises to protect their operations," which may sound reasonable, but only if you know nothing about crypto and security.
The only experimental method to determine whether a security system works is to expose it to public scrutiny, so that peer-reviewers can find the flaws and show you how to patch them ("Anyone can design a security system that they themselves can't figure out how to break" -B. Schneier). The technical term for reliance on "security through obscurity" is "wishful thinking." Using "customized encryption" is like using "improvised brain-surgery techniques."
But even assuming that Barr can repeal the Scientific Method and somehow create reliable knowledge without peer review, he's still got a hell of a problem, which is that you now have to contend with two sources of working cryptographic code: stuff that comes from abroad, and stuff that is in wide circulation in America, and somehow you have to prevent "bad guys" from acquiring that stuff.
But with that all said, Barr is, indeed, less wrong than his predecessors, and as Bruce Schneier points out, that is actually pretty good news, because Barr is at least being honest about the debate. He's not saying that you can ban crypto without making security worse for everyone who relies on it — he's saying that the security gains from banning crypto exceed the security costs.
This gives us an empirical basis for evaluating cryptography bans: we can actually take steps to quantify the costs and benefits of Barr's proposal and weigh them in the balance. My guess is that Barr will then cheat, using motivated reasoning to downplay the risks and overweight the benefits (while still failing to acknowledge that the whole exercise is pointless because "bad guys" will just download and use working crypto from outside of America). But at least other people in the debate will be able to understand its contours better and make up their own minds.
I hope that Barr's latest speech signals that we can finally move on from the fake security vs. privacy debate, and to the real security vs. security debate. I know where I stand on that: As computers continue to permeate every aspect of our lives, society, and critical infrastructure, it is much more important to ensure that they are secure from everybody — even at the cost of law-enforcement access — than it is to allow access at the cost of security. Barr is wrong, it kind of is like these systems are protecting nuclear launch codes.
Attorney General William Barr on Encryption Policy [Bruce Schneier/Schneier on Security]