Dear Catastrophe Robot

Of course we all remember Skynet, the evil internet that uses flying death machines and Arnold Schwarzenegger to ravage the world in the Terminator movies.

At the Technology in Wartime conference this weekend at Stanford, I met the real-world equivalent of Sarah Connor—a group of major-league roboticists, programmers, and computer security geeks who are trying to make sure Skynet never becomes a reality.

Skynet is made up, but the Pentagon has tons of dough and a lot of generals who want to blow shit up. That mix is about as safe as building a meth lab in your garage and to prove it, the US military has put machine guns on robots (though I’m told they don’t intend to use them), built hunter-killer flying bots, and toyed with the idea of launching state-sponsored attacks against enemy cyber-infrastructure.

Of course that’s only the tip of the iceberg. The military brass never seems to lack in twisted creativity when it comes to answering the question “how many ways can you spend $500 billion?” You can be sure plenty more crazy inventions are on the way.

How to deal with that reality is a matter of some debate. Ron Arkin, a robot guy from Georgia Tech, thinks that one day in the not-so-distant future robots capable of making their own life or death decisions will be on the battlefield. Operating under the assumption of inevitability, he dedicates a good portion of his time toward figuring out how to program morals and ethics into gun-slinging automatons.

Arkin’s solution seems to be a mix of programming a robot to follow the Geneva Convention, and developing some way to determine the difference between an innocent and a suicide bomber. He maintains that robots’ lack of self-preservation instinct will help them keep a cool head when the bullets start flying. I wonder how lacking that drive to stay alive wouldn’t convert them into suicide bombers themselves.

I’m singling Arkin out here because I’ve read his work, and because I can’t post on everything from the conference at once. I should say this, though: Arkin’s plan for ethical killer robots may sound little looney, but at least he’s trying. The military isn’t, as far as I know. And most of the folks at the Tech in War conference are taking far stronger stances to try and prevent our armed forces from running amok with their futuristic battle toys.

Good for them, and good luck: they’re going to need it.

Post a Comment

Required fields are marked *


%d bloggers like this: