Close Menu
    Technik

    Anthropic lehnt Anträge des Pentagons im Streit um KI-Sicherheitsmaßnahmen ab, sagt CEO

    Share.
    View 28 Comments

    28 Kommentare

    3. Mr_Greystone on

      We’ve already had a lot of time to become unethical on our own without help from AI.

    7. Saedeas on

      People will surely concede that Anthropic is actually doing the right thing here and not endlessly bitch and make the narrative „everyone is equally corrupt, dude“, right?

    9. angrybobs on

      They really have no reason to accept. Claude is used by so many companies and the revenue from all of them will always be way more than the government. Now the government can either use it or an inferior product. Yeah they can force their supply chain to ban them but I’m sure most will find ways around that.

    11. National-Charity-435 on

      Good for them, then again…secretary or war (against sobriety) kegseth can turn to altman, pichai, and whoever runs pilot

    12. bristow84 on

      You know, if I’m going to support one of these AI companies I’ll at least support the one with a semblance of a backbone.

    14. Shadow293 on

      Bet they’ll cave by tomorrow. Netflix just
      pulled out of the WB deal, so I can’t see this lasting either.

    16. brakeb on

      huh, they do have a standard… maybe someone showed them „The Terminator“ over the weekend.

    17. savagebongo on

      Just tell them no, and if that’s a problem they can move their HQ to europe or asia.

    18. Dalmahr on

      I’m not totally sure how much I trust them after loosening their safeguards the other day. Part of me feels like the combination. Of that and this is to basically be able to say they aren’t letting the military use it this way, they just found a way to do it

    19. Delmoroth on

      Is having grok make these decisions better? That’s what the choice actually amounts to.

      It’s not a choice between autonomous weapons and peace, it’s a choice between smart weapons run by anthropic or smart weapons run by other company’s models. Both of which are likely better than dumb weapons for avoiding civilian deaths, though, a malfunction in the autonomous stuff is pretty dangerous.

      To be clear, I’m not saying this is a good thing, only that it seems to be the situation we are in.

    20. zvoidx on

      *CEO, Dario Amodei*

      His last name has the words „Ai mode“, there’s also „ai“ in his first name. 🤔

    22. Cannabrius_Rex on

      „Regardless, these threats do not change our position: we cannot in good conscience accede to their request,“ Amodei said.

      Thanks for doing the right thing

    23. nigheus on

      Anddd I’ve now switched from ChatGPT to Claude. To date, this is the only big company (aside from maybe Costco) that has actually stood up to fascist wannabe dictator.

    24. Critical-Republic-53 on

      I’ll probably get downvoted for this but— the standard safeguards on these apps make it impossible to use the AI for defense purposes. Prompt: “Help me develop a CONOP that would do xyz”. Reply: “I can’t help with wargaming that’s unethical”. The Pentagon needs versions of these apps that will actually give them answers without the analysts having to argue it out with the AI on the ethics of doing their job. Sure, plenty of other AI companies will give modified versions of their AI, so does the Pentagon need to invoke the Defense Production Act…. no probably not.

    28. Hot-Drummer2191 on

      didn’t we just see a shit ton of articles saying that they ditched these principles? out of the loop here but sounds like it ended well

    Leave A Reply