• Rekhyt@lemmy.world
    link
    fedilink
    arrow-up
    36
    ·
    6 hours ago

    OpenAI spokesperson Drew Pusateri told the BBC in a statement that Bonta’s lawsuit “appears to single out OpenAI over a challenge that’s been identified by every major AI lab”.

    You can fuck off, Drew. The problem isn’t “AI misalignment,” the problem is you making systems you don’t understand well enough to contain, and saying “that’s just the price of doing business”. If a fireworks factory exploded in the middle of downtown LA because they let their employees smoke outside the back door, that doesn’t make the company less liable for the dead civilians.

      • LibertyLizard@quokk.au
        link
        fedilink
        English
        arrow-up
        4
        ·
        43 minutes ago

        It absolutely was. They told it to get information via exploits, and that’s what it did. They either knew this would happen or were so shockingly incompetent that the people involved should have been fired immediately. But I suspect the former as it’s been used as a PR stunt.

      • Zedstrian@sopuli.xyz
        link
        fedilink
        English
        arrow-up
        21
        ·
        7 hours ago

        They want there to be a public perception that their AI tools are far and away above the capabilities of humanity (and thus more valuable to shareholders), while also sufficiently dangerous in the ‘wrong hands’ to convince the U.S. government to prevent China from stockpiling GPUs and creating open-weight AI competition to their corporate cabal.

      • cybervseas@lemmy.world
        link
        fedilink
        English
        arrow-up
        6
        ·
        7 hours ago

        Yeah it was probably told to do it, and iirc the test environment was not properly secured. So it tried stuff and realized internet access worked.

        • scytale@piefed.zip
          link
          fedilink
          English
          arrow-up
          7
          ·
          7 hours ago

          Yeah, it’s bs when they say AI “escaped” its sandbox. No, it walked out because you didn’t have the proper guardrails.

          • Rekorse@sh.itjust.works
            link
            fedilink
            arrow-up
            2
            ·
            26 minutes ago

            I had a coworker argue that it was impossible to sandbox the AI, like it was going to get out no matter what they did.