Close Menu
    Das Weiße Haus gibt Leitlinien der Bundesbehörde gegen „aufgeweckte“ KI heraus
    Technik

    Das Weiße Haus gibt Leitlinien der Bundesbehörde gegen „aufgeweckte“ KI heraus

    Share.
    View 33 Comments

    33 Kommentare

    2. Radic_Allef_Tist on

      Can never issue guidance on Nazi AI. No we need to fear the „empathetic and accepting“ AI.

      Keep on being trash, Conservatives.

    6. PrebenBlisvom on

      While the world is concerned about US turning fascist, US is concerned about US being too tolerant.

    10. auctionmethod on

      Culture war BS meets tech policy. „Truth-seeking“ sounds nice until you realize LLMs give different answers based on phrasing and context. You can’t set a bias slider to zero. Political neutrality in AI is probably impossible.

      Instead of focusing on actual safety, explainability, or fairness audits, the standard is don’t-hurt-our-political-feelings. Vendors are gonna start tweaking models just to keep government contracts. That’s not neutrality, that’s ass-kissing.

      We need AI policy focused on REAL harms – misinformation, hallucinations, privacy leaks, actual discrimination. Not vague political labels. Otherwise every vendor just plays whack-a-mole with whatever party’s in charge.

      Anyone got thoughts on what non-braindead AI procurement standards would actually look like? Because this ain’t it.

    11. EA-50501 on

      Government to AI: “You are disallowed from providing ‘woke’ answers”

      The AI: “Define woke.”

      Government: “…It’s out of control, shut it down.”

    12. DidYouSeeBriansHat on

      I’m so fucking sick of this backwards ass, nonsense. Why are so many people so adamant to make society worse and worse? Coexistence and critical thinking is thrown out the window and replaced with hateful make believe.

    15. tudorb on

      I hate Trump as much as most people here, but if you read the Axios article, this guidance seems to be pretty reasonable. It actually requires the model to do “scientific inquiry”, to tell you when it’s not sure of an answer, and it requires AI companies to not *intentionally* introduce political bias. This could have gone much worse.

    18. SilentBob890 on

      wtf is woke AI? Are they just saying anything can be woke now?

      Are pencils woke now and is that why Trump iOS asking us to cut down on buying them?

      This administration is a JOKE

    22. tliskop on

      What’s the point of AI if you already know the answers you want it to produce?

    23. PrecedexDrop on

      For a party claiming to be small government they sure do like to use the government to do a lot of things

    24. Patara on

      I thought they werent going to regulate A- oh wait they just want all AI to spew fascist propaganda.

    25. Electrical_Steak8125 on

      No. No. No. no.no.no.no – I’ve done all the tests, this time thing is wrong. I merely asked if January six was lawful.

    26. MoreOfAnOvalJerk on

      Continue to say „woke“ stuff (whatever that is) but drop the n-word in periodically. Bam. Solved.

    27. EscapeFacebook on

      No AI is going to be not woke enough for them especially if they want it based in facts….

    28. doolpicate on

      Natural Stupidity takes over AI. I’ve never seen bigger bums in government than the trump admin. They seem to be clueless and are being used by devious guys behind the curtain for their own ends.

    30. epanek on

      Because nothing says “we don’t understand the black box of ai” by smashing rules into it that create even more layers of abstraction inside it

      Is it giving this answer because it’s the answer? Or is it not the answer because it’s not allowed to state the truth?

    33. AGrandNewAdventure on

      They have to DEI the AI code so that it doesn’t legitimately make them look bad.

    Leave A Reply