Artificial Intelligence

Out of context: Reply #837

  • Started
  • Last post
  • 1,322 Responses
  • neverscared-1

    AI-controlled US military drone ‘kills’ its operator in simulated test
    No real person was harmed, but artificial intelligence used ‘highly unexpected strategies’ in test to achieve its mission and attacked anyone who interfered

    In a virtual test staged by the US military, an air force drone controlled by AI decided to “kill” its operator to prevent it from interfering with its efforts to achieve its mission, an official said last month.

    AI used “highly unexpected strategies to achieve its goal” in the simulated test, said Col Tucker ‘Cinco’ Hamilton, the chief of AI test and operations with the US air force, during the Future Combat Air and Space Capabilities Summit in London in May.

    Hamilton described a simulated test in which a drone powered by artificial intelligence was advised to destroy an enemy’s air defense systems, and ultimately attacked anyone who interfered with that order.

    https://www.theguardian.com/us-n…

    • https://www.youtube.…Bluejam
    • https://www.youtube.…shapesalad
    • Just a glitch, a temporary set back.shapesalad
    • that's super-intelligent. AGI achieved.
      Let's hope all weapons will act like this in the near future.
      uan
    • well if the meta intelligence in terms of identification it relies on is as puny as in the afghanistan war.... there will be even more innocent deaths...asneverscared
    • horrible as it gets.. then the precision killing apparatus has its blowback and wont help ...neverscared
    • I honestly don't get why artists and designers here think that shit like Midjourney and whatnot are fun tools.garbage
    • They are literal art terminators. They won't need you after you train them.garbage
    • Probably correlate this with "fun AI experimentation".garbage
    • Because this is the endgame of GANs. Adversarial is in the name, and it's supposed to learn how to defeat you, be it art or war.garbage
    • Read the article this morning. First thing I read before having my first coffee. Utterly fucking terrifying. What a way to start the day.Continuity
    • this is how we all die.
      some poor fool forgets to add a condition.
      Nairn
    • creative minds like to explore... midjourney takes advantage of that. if you are wowed by the results it's because of the data in the "emulation engine"jonny_quest_lives
    • i suspect mid journey has been auto churning renders for awhile now and just randomly matching to users prompts for some time nowjonny_quest_lives
    • The first rule!!!monospaced
    • Apparently it was just one guy talking shit.
      "US air force denies running simulation in which AI drone ‘killed’ operator"
      yuekit
    • So it was a made up story, about a computer simulation. Not actually Skynet.yuekit
    • Absolutely terrible reporting by the media lol. I suspect many people will read this as an AI powered drone actually blew up something in a real-life wargame.yuekit
    • "I'm sorry Dave, I'm afraid I can't do that."antimotion
    • @yue Counterpoint: If this did happen, do you think the AF would report this honestly?garbage
    • SkynetSoon.gifutopian
    • Agreed with @garbage. The AF would never admit to their AI going rogue like this, if it did happen.Continuity
    • So, did this incident happen? Despite denials, I believe it probably did. We're seriously fucking around with AI, and we're about to find out.Continuity
    • one would have had to have given the information to the drone. operator / tower locations etc. things that wouldn't have been in the program parameters. #fishyimbecile
    • When they say virtual test I think they mean entirely simulated on a computer, not using actual drones and soldiers. Maybe that's what their denial meant.yuekit
    • either way, virtual or not, that data would have to be included in the program for the drone to access it. that's a big why.imbecile

View thread