The FBI’s Face Recognition Project Isn’t Scary

One consistent problem with conspiracy theories is that they almost inevitably center on government being absolutely incompetent in every respect except law enforcement. Every law enforcement body in America is not just a well-oiled machine, it’s a cobra, stalking our civil rights and privacy through the grass, waiting to munch on them like a rat or something.

The idea that for most of these people it’s genuinely just a job and maybe not one they do particularly well doesn’t fit in with the whole conspiracy-thriller idea, possibly because it’s much scarier that these people really are trying to violate our civil rights for “the greater good” and are too incompetent to even use their illegally acquired tools effectively.

Which brings us to the current Internet freak-out over the FBI’s Next Generation Identification program.

It’s not wrong to be worried about the FBI violating the civil rights of Americans: They do it way too often. But, first of all, they’re not very good at it. They get caught a lot. And secondly, people are mistaking ambition with ability.

The idea behind the Next Generation Identification program is undeniably ambitious; essentially it would put all identifying marks on a criminal, from fingerprints to iris identification, in one place. You might be wondering what would keep, say, an unethical FBI agent from tracking a private citizen not suspected of any crime outside of internal protocols. Not much, really, at least not in what the FBI has laid out, and that’s a problem.

The Internet, though, seems convinced that because the government is throwing a billion dollars at it, and we have lots of HD cameras everywhere, that this is a bullet in the head of privacy. The FBI will be able to track everybody! There will be no privacy! Big Brother has arrived!

Uh, no. Let’s start with the most basic problem with that statement: The FBI has no access to a good chunk of the cameras you see. No, seriously. Most law enforcement camera networks are centered around public transit, high-traffic public areas, and roadways. It’s true there are more and more camera networks, and that’s an issue that gets too little discussion. But your average drug store surveillance camera is being watched by a bored ex-cop looking for shoplifters.

Secondly, the FBI may have grand ambitions of tracking people with facial recognition software, but the reality of the technology is far short of Enemy of the State. It’s true that hats and sunglasses can’t fool current systems, but A) they have to find you first, which takes a lot of processing power, and B) they have to care where you are. Most facial recognition software out there uses a static image and compares it to another static image. Doing that with live video is a much trickier and resource intensive process.

Thirdly, this assumes that the system will get this far, work as advertised, and be properly applied. This is not a safe assumption. The government is very, very bad at commissioning, building, and using computer systems. The FBI is no exception: For years, there were rumors of a system to intercept all emails code-named “Carnivore” that got destroyed because nobody in the FBI had heard of spam mail.

The simple fact of the matter is that if the FBI were actually good at this stuff, we’d never know about it. Instead, the Electronic Frontier Foundation is busting their chops on a regular basis. These are people who cut down the wrong door with a chainsaw and let the real criminals escape. The FBI is run by people, and people make mistakes. They make a lot of them.

Is there the potential for abuse here? Absolutely, and the FBI hasn’t done nearly enough to address what accountability there is or how agents will be called to account if the system is misused. On the other hand, it’s also worth noting that we’re not building Skynet, here, either. Just perhaps giving another form of chainsaw to the wrong person.