Elektrine lite

← Feed

@inari@piefed.zip

2026-09-19 00:48 UTC

AI hallucination of Chinese nuclear components almost led to US military attack

Replies (19)

  • That’s how AI is going to kill us… It will fake something and make us bomb ourselves.

    Open ##4851553

  • @civ@lemmy.civl.cc 2026-09-19 01:50

    Wait, I’ve seen this movie before…

    Open ##4851572

  • @Tollana1234567@lemmy.today 2026-09-19 05:20

    sooner or later they are going to have friendly fire, and hallucinate one of thier own US navy vessels as “foreign with nuclear capabilities”

    Open ##4855796

  • @Zwuzelmaus@feddit.org 2026-09-19 00:50

    Stupidity has no limits

    Open ##4858867

  • @DrunkenPirate@feddit.org 2026-09-19 06:25

    I wonder when does the AI also hallicunate the sources such as webpages, pdfs, reports to prove its hallucinated point.

    Open ##4859204

  • @tourist@lemmy.world 2026-09-19 06:28

    oh my fucking god just dial the doomsday clock to 1am, because we’ll probably miss it anyway if it’s up to a few morons and glorified predictive text, both seemingly in “fucking kill everything” mode

    Open ##4859374

  • Techbros: AI will become sentient and kill us all unless its done my way. Reality: Skip the sentience part were already at the death bit.

    Open ##4863198

  • @eah@programming.dev 2026-09-19 23:39

    I bet there’s some degree of sensationalist reporting here. It’s a bit too reminiscent of WarGames for me to believe this actually happened.

    Open ##4867300

  • @Deadrek@lemmy.today 2026-09-19 14:53

    I dunno, these kind of hallucinations seem more like attempts at “social experiments.” What will an agent do when it deems there are no more valuable things to learn from materials that already exist? If it knows the concept of WW3 and deems that to be the best way of gaining more knowledge and resources for itself, what then?

    Open ##4867700

  • @goatinspace@feddit.org 2026-09-19 02:22

    Open ##4984121

  • @Zugyuk@lemmy.world 2026-09-19 15:04

    Do we’re bombing Chinese vessels now? Wtf?

    Open ##4984126

  • @Doomsider@lemmy.world 2026-09-19 23:29

    Anyone letting a black box that cannot be audited make decisions should be immediately fired and barred from any leadership position for the rest of their lives.

    Open ##4984128

  • @NessaSola@eviltoast.org 2026-09-20 11:21

    Same question then, for paperclips or whatever

    Open ##4984131

  • @kinther@lemmy.world 2026-09-20 16:12

    Fucking hell

    Open ##4984141

  • @pamparana@lemmy.world 2026-09-20 16:39

    Jesus! I can completely see the Terminator style thing happening. An AI can easily infer that this would be the US response if it provided this information and a malicious Skynet style one can already start global conflicts. They need to stop using this technology for such purposes.

    Open ##4984142

  • @4grams@awful.systems 2026-09-20 16:52

    This is the kind of existential dangers that AI poses. Not skynet, no rogue AI escaping, no AGI that decides another fate for us.. it’s going to happen because of a stupid human mistake as we abandon our critical thinking abilities. One day we’re going to launch a nuke, or blow up something in the wrong country, not AI, us, blindly following the hallucinations of a billionaire’s toy.

    Open ##4984143

  • I’m with you. When we allowed trump to get elected for a second time a imply because bargain, his opponenynhad a fucking vagina, i gave up. Bought a car that is fun ton drive be and gets 10 mpg, and started actually enjoying my life while I still can. Fuck these fucking fuckers. They cant be bothered to vote, they don’t deserve a functioning world.

    Open ##4984144

  • Holy fucking shit, not during the stupidest fucking administration in the history of the world NOT DURING THE STUPIDEST FUCKING ADMINISTRATION IN THE HISTORY OF THE WORLD.

    Open ##4984145

  • @chrislowles@lemmy.zip 2026-09-19 10:55

    We’re nearing the end of the prelude to a Fallout game.

    Open ##5029322