Sandboxed experiment found itself a zero day, escaped onto the open internet and validated scary predictions about rogue agents
  • BestBouclettesEnglish
    6·
    2 months ago

    If you’re responsible for any code generated by an AI, you should also be responsible for any action taken by your AI agent.

    • prettybunnys@piefed.socialEnglish
      2·
      2 months ago

      In the same way parents ought to be responsible for their children, or pets, excepting those things are actually snetient