lol like maybe some technocrat will be president and his vice president is a machine and he resigns so the algorithm can control the country and it just immediately nukes everything or makes crime extra illegal and lock everybody up or some shit. i'm sure some sci-fi book has played with this idea or something, just seems like an actual thing that could happen only because it's so insanely stupid it would make sense it's what would happen.

  • crompo [none/use name]
    ·
    4 years ago

    In all seriousness AI definitely won't be at a point to do that any time soon. I do think that AI might some day be used to generate policies that actual human politicians then vote on. And it'll be heralded as objectively good policy because it came from a machine that "did the math", ignoring the reality that AI will always operate under the biases of the people who programmed it.

    man I just made myself sad about a fake thing I made up

    • 5bicycles [he/him]
      ·
      4 years ago

      There's thinktanks out there already heralding this type of bullshit as the end-all-be-all solution to things like human bias'. You'll get your AI that does things but technically somewhere in the process there's a human who has to click "OK" and that makes it fine by lib-logic.

      That was the solution I was last presented at work over this type of nonsense. The tech does all the decisions, but you just put a human in there as a fall guy when inevitably the algorithm suggests to kill all the poor so you have an individual to prosecute. Of course, on paper all the OK-clickers are made to sign forms that they definitely understand how the algorithm or AI works because obviously that's a skill that's very common among elected officials and public servants and that solves the one problem AI if you don't actually care about any of the consequences: Who do we throw in jail?

    • JoesFrackinJack [he/him]
      hexagon
      ·
      4 years ago

      lol yeah, i do think algorithms will play a bigger role in day to day things. Like for example i'm sure they're used already in some things like insurance industry or medical stuff maybe? with human supervision but i can see an algorithm deciding when or who they approve insurance claims for based on some criteria. it would save the company so much money cutting staff. Barbaric to think about though tbh