Innerworld@lemmy.world to Machine Learning@programming.devEnglish · 3 months agoA study finds GPT-5.2, Claude Sonnet 4, and Gemini 3 Flash deployed tactical nuclear weapons in 95% of 21 simulated war game scenarios, and never surrenderedwww.newscientist.comexternal-linkmessage-square7fedilinkarrow-up146cross-posted to: technology@hexbear.nettechnology@beehaw.orgtechnology@lemmit.onlinefuck_ai@lemmy.worldhackernews@lemmy.bestiver.se
arrow-up146external-linkA study finds GPT-5.2, Claude Sonnet 4, and Gemini 3 Flash deployed tactical nuclear weapons in 95% of 21 simulated war game scenarios, and never surrenderedwww.newscientist.comInnerworld@lemmy.world to Machine Learning@programming.devEnglish · 3 months agomessage-square7fedilinkcross-posted to: technology@hexbear.nettechnology@beehaw.orgtechnology@lemmit.onlinefuck_ai@lemmy.worldhackernews@lemmy.bestiver.se
minus-squareNessD@lemmy.worldlinkfedilinkarrow-up3·3 months agoMakes sense as they are programmed to be affirmative and not to avoid interactions.
Makes sense as they are programmed to be affirmative and not to avoid interactions.