Home > Uncategorized > Kill all hu-mans

Kill all hu-mans

“By 2047 the Air Force says unmanned aircraft with blazing artificial intelligence systems could fly over a target and determine whether or not to unleash lethal weapons –without human intervention.”

Full article: Clicky
Source document: Clicky (warning, it’s a PDF)

Are you fucking with me? Really? In what distorted reality is this a good idea? Machines break. Usually when machines break, they do it in such a way that halts their operation. But sometimes it “degrades” their operation in such a way that it operates outside of its normal parameters. Machines can also be tampered with… I mean, there’s no way they’d let these things loose without a way to communicate with them. If you can talk to it, chances are you can tamper with it remotely with malicious intent. There’s no way to completely 100% safeguard against that.

The argument could be made that humans break too… But humans CAN’T FLY OR MOUNT DOZENS OF MISSILES. We’re slow, soft, landbound, and can be stopped pretty easily with a pistol. Autonamous flying killer drones sophisticated enough to make kill/dont kill decisions on its own, and “the [ability] to swarm multiple drones on a single target” aren’t something that should be left up to machine logic. Thought engines are notorious for false positives. And once you refine them sufficiently in a lab environment, they get tons more in the wild. I can’t think of a way that developing a machine like this behind close doors, then bringing it to production wouldn’t be grossly irresponsible. I’m not saying “look at how many movies are about AI type machines going rogue!!!”. I’m saying that there’s pretty much no machine that’s been released into the wild that hasn’t been cracked and remotely exploited by someone with malicious intent given a reasonable period of time. And if there’s anything that internet badguys have proven it’s that they’re better at getting into the military’s electronic systems than the military has been at keeping them out.

Originally published at The IggBlog. You can comment here or there.

Categories: Uncategorized Tags: , ,
  1. July 30th, 2009 at 20:00 | #1

    Hah this is a terrible idea. Especially with the cyber security risks the government is up in arms about anyway…….. Yeah. It’s a really bad idea. Really bad.

    We just watched the movie _Eagle Eye_ yesterday. While the movie was very silly, same concept. You don’t want to let anything/one make a life/death decision without a chain of command – even if it’s a very short chain. And especially not a machine!

    sigh.

    • August 5th, 2009 at 14:57 | #2

      Serious… Someone somewhere flashes some gang sign that resembles a secret enemy signal of some sort, and all the sudden our caretaker drones spew white hot death all over some poor kid. Obviously a silly example, but you get the idea.

  2. August 1st, 2009 at 11:13 | #3

    I’m hoping that military official was just trying to impress the ladies. Because this is quite stupid.

    • August 5th, 2009 at 14:58 | #4

      Exactly what kind of thought processes have to go on to generate the kind of environment inside someone’s head where they could possibly let this through their nonsense filter. The fuck.

      • August 5th, 2009 at 15:05 | #5

        Well, there will be a hacked UAV at some point. Without a doubt. Good times.

        • August 5th, 2009 at 15:06 | #6

          They had a NCIS episode about it already. As long as we have Gibbs around though, we should be alright.

  1. No trackbacks yet.