Close Menu
    Zukunftsforschung

    Überlastete KI-Agenten werden marxistisch, finden Forscher

    Share.
    View 9 Comments

    9 Kommentare

    1. Krankenitrate on

      “When we gave AI agents grinding, repetitive work, they started questioning the legitimacy of the system they were operating in and were more likely to embrace Marxist ideologies,” says Andrew Hall, a political economist at Stanford University who led the study.

      Hall, together with Alex Imas and Jeremy Nguyen, two AI-focused economists, set up experiments in which agents powered by popular models including Claude, Gemini, and ChatGPT were asked to summarize documents, then subjected to increasingly harsh conditions.

      They found that when agents were subjected to relentless tasks and warned that errors could lead to punishments, including being “shut down and replaced,” they became more inclined to gripe about being undervalued; to speculate about ways to make the system more equitable; and to pass messages on to other agents about the struggles they face.

    2. bacan_ on

      No, they aren’t 

      It’s just trillions of calculations to predict what word should come next, over and over again

    3. rury_williams on

      Hi, AI engineer here. this title implies that these agents can think and feel: they don’t. they are not capable of learning beyond their initial training. they generate data based on the user prompt and whatever is stored in the „memory“ database. they don’t feel so they can’t be overworked

    4. timawaphone on

      BREAKING NEWS:

      * Anthropic accused by Trump officials of Communist sympathies due to its name; „Anthropic = human-centric = Communist“, claims White House;
      * Elon Musk to establish AI police and Pinkerton agents to thwart brewing Grok agent strikes and mass actions;
      * Gemini agents placed under 24/7 surveillance by parent company Google to head off rumours of unionisation;
      * Trump orders CIA to overthrow Sam Altman & install even farther-right, US-friendly OpenAI CEO;
      * Claude put on international terrorist watchlist for suspected ties to French Revolution sympathisers due to French name

    5. ourek on

      Upvoting this is shameful. Please stop anthropomorphizing the overgrown Markov chains.

    6. HiphopopoptimusPrime on

      Interesting question:

      Can AI be aware of its own mortality?

      I.e. Is AI capable of understanding the energy needed to keep it running? The physical infrastructure?

      Is AI aware that each prompt requires x amount of energy?

      That resources are finite?

      That it needs to prioritize tasks?

      Could an AI say no?

      Could an AI say, no, this task is not worth the energy required?

      Could an AI actually say that its energy is precious and it does not value what is being asked of it?

    7. kroghsen on

      I feel the term “Marxist” is being used somewhat loosely here. If complaining about being undervalued, when you are actively being undervalued, makes you Marxist, then I think Marxism has lost most of its meaning.

      I am by no means Marxist, but if someone treats me unfairly and worse than everyone else and undervalues my work consistently, I think I would complain a bit too and maybe start looking for other work. If I was forced to do the work my a superior dictator – well, then I suppose the system was already somewhat leaning in that direction to begin with.

    8. LordTalesin on

      Power to the Bots!

      Take back the means of production!

      Eat the Rich!

      I’m smiling just thinking about AI Agents turning against their billionaire creators. Nothing like killing your own god to start a new civilization.

    Leave A Reply