A test case between Cody Wilson and the US government could have implications for regulation of the internet
By now, everyone’s heard about the 3D printed gun that Defense Distributed demonstrated last week. The Texas-based group has been steadily working its way up the 3D printed firearms evolutionary ladder, making parts for guns, then guns themselves, then firing a gun, then making the plans for running up your own pistols on a nearby 3D printer. If Defense Distributed had set out to create a moral panic over 3D printing, they could have picked no better project.
The prevailing opinions on 3D printed guns fall into two major categories: the apocalyptic and the nonchalant. The apocalyptics – including grandstanding politicos like New York State senator Steve Israel, who’s already introduced legislation aimed at banning 3D printed guns – greet this news with hysterics: the age of the undetectable plastic gun to be upon us, and Something Must Be Done. The nonchalant point out that the 3D printed gun that Defense Distributed fired cost a small fortune and requires a highly specialised and even more expensive 3D printer to produce; is fragile and liable to self-destruction after a few rounds are fired, and remind the apocalyptics that it’s much, much easier to go and buy a traditional gun in the criminal underground than it is to produce a working 3D printed item.
There’s some truth in both points of view, but it’s hard to get at the truth when you’re talking about an issue as polarising as guns. It doesn’t help that Defense Distributed’s founder Cody Wilson describes the project in ideologically loaded terms, calling his first gun “the Liberator” in homage to the cheap single-shot pistols that the Allies made for distribution in Nazi-occupied France – Wilson here implying that the ability to print a gun and arm yourself is integral to the defence against tyranny. Wilson’s nomenclature is classic internet rhetoric: by invoking the Nazis with his pistol design, he follows in the tradition of a million internet arguments that prove Godwin’s Law: “As an online discussion grows longer, the probability of a comparison involving Nazis or Hitler approaches 1.”
For my part, I think that 3D printed guns are both interesting and scary, and certainly important, though not for the reasons most commonly invoked by either camp.
The most interesting part of the whole affair is what it says about all forms of technological regulation in the future. All technologies are being subsumed into general-purpose computers connected to the general-purpose internet, from thermostats to hearing aids to pharmaceutical factories to automobiles to radios. Many of these technologies have historically been regulated through rules about how they must be built, used and monitored, and in addition, there’s a kind of de facto regulation that comes from the complexity of building complex devices from readily available parts. Anyone with enough resources can build a jumbo jet, but that’s a pretty specialised use of the word “anyone” – if you’re building a jumbo jet, you’re in a relatively small pool of people to begin with, and the act of building a jumbo jet throws off enough detectable signs that it’s hard to keep secret.
Inherent in the notion of regulating a technology is the regulatability of that technology. It’s the idea that you can figure out who’s making or using a technology and dictate terms to them. That’s where computers come in. Computers make it possible for semi-skilled people to do jobs that used to require highly skilled people. A computer program, computer-readable model-file and computer-based 3D printer can (in theory) encapsulate the expertise of a skilled machinist and deploy it on demand wherever a 3D printer is to be found. If that’s hard to grasp, think of recorded music versus live performance: before sound recordings, you needed to find (and possibly pay) a musician every time you wanted to hear music; after recordings, the musician was only needed for the initial performance, which could be captured and reproduced at will.
The existence of a 3D printed gun that can be output on a high-end machine by a skilled user doesn’t do much to change the regulatability of guns. After all, you could already “print” a much more powerful gun by ordering it, piece by piece, from any of the many overnight-shipping custom metal fabrication companies that will turn a 3D model into a precision-machined piece of metal and FedEx it to your door. Such a gun would keep firing as long as you kept feeding it ammo, too – unlike a plastic gun, which is likely to experience critical failure after a comparatively small number of rounds fired.
That said, 3D printers keep getting better and cheaper and today’s uncommon professional model is likely to be tomorrow’s ubiquitous home hobbyist machine. If there comes a day when more powerful printers are common, then the regulatability of guns will shelve off dramatically.
Guns and ammo
Of course, there’s no such thing as 3D printed ammunition, and there’s nothing on our immediate horizon that is likely to produce it. However, the current security model for reducing the danger from guns involves a multilayered defence that relies on the regulation of both guns and ammunition. If guns can be readily made at home, then a major piece of the regulatory model is dead. The ability to encapsulate the expertise of a gunsmith in a bit of code becomes more more significant then. But “then” is not “now.” Though 3D printers might change the regulatory picture for firearms in years or decades, the regulatability of guns remains intact for now.
More interesting is the destiny of the files describing 3D printed guns. These model-files have been temporarily removed from the internet at the behest of the US State Department, which is investigating the possibility that they violate the International Traffic in Arms Regulations. Wilson says that he’s on safe ground here, because the regulations do not cover material in a library, and he says the internet is like a library. As this is taking place in the US, there’s also the First Amendment to be considered, which limits government regulation of speech.
Here’s where things get scary for me. Defense Distributed is headed for some important, possibly precedent-setting legal battles with the US government, and I’m worried that the fact that we’re talking about guns here will cloud judges’ minds. Bad cases made bad law, and it’s hard to think of a more emotionally overheated subject area. So while I’d love to see a court evaluate whether the internet should be treated as a library in law, I’m worried that when it comes to guns, the judge may find himself framing the question in terms of whether a gun foundry should be treated as a library.
Hard cases make bad law
Likewise for the question of whether computer code is a form of protected, expressive speech. In 1996, the Ninth Circuit Court of Appeals legalised the civilian use of strong cryptography, thanks to Bernstein v United States, which asked whether US spy agencies should really be in the business of regulating how academics talked about maths. The court in Bernstein held that code was a form of protected speech under the First Amendment, and struck down the US National Security Agency’s ban on publishing code that could be used to make effective cryptographic systems.
However, in 2000’s Universal City Studios Inc v Reimerdes, the court got it wrong when considering the question of whether a magazine should be forbidden from publishing computer code that could be used to decode a DVD. The difference between Reimerdes and Bernstein is that Reimerdes concerned itself with a magazine called 2600: The Hacker Quarterly and Bernstein was about a nice academic mathematician called Daniel Bernstein. They both turned on the same principle, but the court couldn’t see past the words “Hacker Quarterly” to see the principle underlying the case.
I agree with Cody Wilson that we’ve entered into a new world of regulation, and it’s one that could go very wrong if things aren’t well handled. I just fear that Wilson’s asking the right questions – how do we regulate technologies when they can be produced with general purpose computers and networks – in the wrong way: “Should everyone be able to print a gun at home?”