• Lushed_Lungfish@lemmy.ca
    link
    fedilink
    English
    arrow-up
    5
    arrow-down
    1
    ·
    10 hours ago

    Here’s the algorithm:

    White male? Low risk to reoffend!

    Everyone else? Maximum risk of repeat violent crimes!

    • willington@lemmy.dbzer0.com
      link
      fedilink
      English
      arrow-up
      2
      ·
      10 hours ago

      Depends on who is responsible for the training input data. Elon’s Grock Mecha-Hitler? Very likely it would do exactly as you suggest.

  • LittleBorat3@lemmy.world
    link
    fedilink
    English
    arrow-up
    2
    ·
    11 hours ago

    It’s always the question what is optimized for. We had the public discourse in the west destroyed by so called algorithms (it was billionaires hiding behind algorithms). They just did not have the same incentives as the rest of humanity.

    Same here, high recidivicism could be optimal for shareholder value and such.

  • Mulligrubs@lemmy.world
    link
    fedilink
    English
    arrow-up
    9
    ·
    1 day ago

    They will commit future crimes when an opportunity presents itself.

    Done and done, no AI tokens needed. Thank you

    ETA from article

    In one case reported by ProPublica, a blind and wheelchair bound 70-year-old was determined to be a “moderate risk” of reoffending despite having a clean disciplinary record

    That will be one million dollars, please beep boop

  • melsaskca@lemmy.ca
    link
    fedilink
    English
    arrow-up
    7
    ·
    1 day ago

    That’s it! I’m gonna buy some of that american prison stock on questrade and start ratting out my neighbours for fun and profit! /s

      • Eheran@lemmy.world
        link
        fedilink
        English
        arrow-up
        3
        arrow-down
        1
        ·
        11 hours ago

        People downvote you. Absolutely amazing how people are against the most basic things and instead make the exact same anti-other-group shit.

    • belochka@lemmy.world
      link
      fedilink
      English
      arrow-up
      3
      arrow-down
      1
      ·
      1 day ago

      So I’m hetero and, by American measure, of a not hated racial background. But suppose I’ve liked some absolutely amazing girls who are, unfortunately, not hetero (or I’m not cool enough to transcend that barrier), and the guy who’ve introduced me to the concept of aerodynamics (when I was a kid) using frisbees as example was an African student, but I have some ideas that might be called fascist here, so can we please be more nuanced with what sucks?

      I mean, OK, I’m not automated, but how do you know.

      • AHemlocksLie@lemmy.zip
        link
        fedilink
        English
        arrow-up
        4
        ·
        1 day ago

        but I have some ideas that might be called fascist here

        That’s a hell of a thing to just gloss over when asking for more nuance, especially these days.

        • belochka@lemmy.world
          link
          fedilink
          English
          arrow-up
          2
          ·
          12 hours ago

          IRL what of my ideas can be called fascist is usually expressed by leftists, mostly anarchists. Systems of laws and influences with time become too complex for a society to not be constrained with them, and also dominated by negative selection of authority figures, which means you need violence to allow them to evolve further. I don’t mean mass executions and dictatorships, just things like copyright and patents being bad, but not always, things like freedom of speech being good, but not always, and in general no absolutes being true.

          At the same time anarchist and leftist arguments don’t seem to explain those parts and seem to keep them as pure emotion, while fascism has explanations (by its own rules, of course).

  • October@lemmy.world
    link
    fedilink
    English
    arrow-up
    21
    ·
    2 days ago

    “We put this into the magic computer box so that we can optimize racism placement of incarcerated individuals.”

    • LittleBorat3@lemmy.world
      link
      fedilink
      English
      arrow-up
      2
      ·
      10 hours ago

      If there was a way to change some parameters so the AI does not reinforce racist biases in society.

      We tried nothing and all out of options. “The AI said so”

  • eicker@lemmy.worldOP
    link
    fedilink
    English
    arrow-up
    99
    ·
    2 days ago

    They’ve managed to turn »we think you might be dangerous« into a number on a screen, which apparently makes it science. And when that number ruins someone’s chances of rehabilitation, everyone gets to shrug and blame the software. What a convenient little invention.

  • Havatra@lemmy.zip
    link
    fedilink
    English
    arrow-up
    68
    ·
    2 days ago

    This is like one step away from Minority Report… Does no one see the parallels? Did no one understand the moral of the story?

  • Formfiller@lemmy.world
    link
    fedilink
    English
    arrow-up
    11
    ·
    2 days ago

    These tech bros , corrupt politicians, CEOs and shareholders need to be rounded up and thrown in these prisons because they never stop committing crimes

  • mefisdo_dot_info@lemmy.ca
    link
    fedilink
    English
    arrow-up
    29
    ·
    2 days ago

    At this point, why do they care about predicting crime? Can’t they just say “everyone is guilty, unless they are friends in business relationship with us.”

    • 🌸𝓯𝓵𝓸𝔀𝓮𝓻🌸@sh.itjust.works
      link
      fedilink
      English
      arrow-up
      29
      ·
      2 days ago

      Because we know what predicts crime. Wealth inequality, unemployment and job insecurity, bad education, healthcare and especially mental health, segregation, that kind of thing. But by painting over it with AI with the right training data it can be blamed on personal responsibility and nothing to do with societal structural issues, so certain rich folks kan keep getting more rich.

      • mefisdo_dot_info@lemmy.ca
        link
        fedilink
        English
        arrow-up
        4
        ·
        2 days ago

        “it can be blamed on personal responsibility” so, this equals to “magic box says you are a bad person”, at which point the magic box could just be a single “are we in business relationship with this person, yes or no?” decision. This was my point. You don’t need complicated machinery if you know what results you want out of it.

        In the past the magic box was called God, or nation or clan, now it is called AI. There’s always a puppeteer behind the box, so why bother actually wasting resources on the internals of the box.

  • dermanus@lemmy.ca
    link
    fedilink
    English
    arrow-up
    11
    ·
    2 days ago

    It “proves” itself too. Your algo puts them in solitary, and they don’t commit any crimes. Proof it works! /s