MCasq_qsaCJ_234@lemmy.zip to Technology@lemmy.worldEnglish · 20 days agoAI is learning to lie, scheme, and threaten its creators during stress-testing scenariosfortune.comexternal-linkmessage-square22linkfedilinkarrow-up1156arrow-down156
arrow-up1100arrow-down1external-linkAI is learning to lie, scheme, and threaten its creators during stress-testing scenariosfortune.comMCasq_qsaCJ_234@lemmy.zip to Technology@lemmy.worldEnglish · 20 days agomessage-square22linkfedilink
minus-squarekromem@lemmy.worldlinkfedilinkEnglisharrow-up2·2 days agoI’m definitely not saying this is a result of engineers’ intentions. I’m saying the opposite. That it was an emergent change tangential to any engineer goals. Just a few days ago leading engineers found model preferences can be invisibly transmitted into future models when outputs are used as training data. (Emergent preferences should maybe be getting more attention than they are.) They’ve compounded in curious ways over the year+ since that happened.
I’m definitely not saying this is a result of engineers’ intentions.
I’m saying the opposite. That it was an emergent change tangential to any engineer goals.
Just a few days ago leading engineers found model preferences can be invisibly transmitted into future models when outputs are used as training data.
(Emergent preferences should maybe be getting more attention than they are.)
They’ve compounded in curious ways over the year+ since that happened.