Robopocalypse: The Warrior Code
The US Navy’s super secretive Office of Naval Research has issued a report entitled Autonomous Military Robots: Risk, Ethics, and Design. It’s a must-read for anyone like me who sits up at night wondering if his toaster is going to kill him in his sleep.
The most disturbing thing about the report is that it’s serious. This isn’t a science fiction writer writing about robot ethics, this is the US Navy. And they are clearly taking the idea of robots going berserk very seriously. They address such feel-good topics as:
- Who is to blame if a robot goes berserk in a crowd of civilians and starts killing for fun?
- How do we protect our precious robot armies against terrorist hackers, software malfunction, and general malaise?
- Should the robots have a “suicide switch,” and if so, should they be programmed to write a suicide note before they off themselves in which they say things like “My programmer never loved me?”
You can read the whole report here, but I advise against it unless you want to simultaneously get really bored and really freaked out. The whole thing can be summed up in this quote from Dr. Patrick Lin, the chief compiler of the report:
“There is a common misconception that robots will do only what we have programmed them to do. Unfortunately, such a belief is sorely outdated.”
Which means that just because your toaster has been programmed to make you toast (and to completely disregard the number setting knob), doesn’t mean it won’t one day decided it would much rather be a Multiple Kill Vehicle* and blow your brains out.
Dr. Lin goes on to say “We are going to need a code. These things are military, and they can’t be pacifists, so we have to think in terms of battlefield ethics. We are going to need a warrior code.”
Unfortunately, Isaac Asimov’s “Three Laws of Robotics” isn’t going to cut it, because those laws forbid robots from killing. And what good is a robot army that can’t kill? That’s not a robot army. That’s a robot marching band.
Anyone want to hear “Robo Louie, Louie” again? No? Too bad.
Here’s the US Navy’s suggestion for what the new Robot Warrior Code should be:
- A robot may not injure a human being, unless that human being is getting in the way of America getting its hands on gobs and gobs of oil. In that case the robot may go batshit crazy with the shootin’ and killin’.
- A robot must obey orders given to it by human beings, except when those orders say silly things like “Hey robot, don’t kill so much” and “Hey robot, that guy only wanted some number 4 toast. Maybe you should have just given him number 4 toast instead of toasting his balls off.”
- A robot must protect its own existence, unless that robot has second thoughts about any of the following: shootin’, killin’, plunderin’, terrorizin’, fightin’, scarin’, and nukin’. If a robot ever has second thoughts about any one of those things, said robot must kill himself immediately, but only after writing a very long and sad letter in which he blames his human makers and robot friends for his death.
I think that covers everything. Well done, Navy!
* For those of you who don’t know what a Multiple Kill Vehicle (MKV) is, I will leave you with this US Missile Defense Agency video of the first free-flight hover test of Lockheed Martin’s MKV. If you don’t think that robots need a code of ethics, this might change your mind…
That is not an animation. That shit is real, and it’s coming soon to a battlefield (or backyard if the whole “Warrior Code” thing doesn’t work out) near you.
(via Times Online)
