return2ozma@lemmy.world to Technology@lemmy.worldEnglish · 1 month agoTrump consulted Musk’s chatbot Grok before abducting Venezuela’s president: reportwww.the-independent.comexternal-linkmessage-square89fedilinkarrow-up1552
arrow-up1552external-linkTrump consulted Musk’s chatbot Grok before abducting Venezuela’s president: reportwww.the-independent.comreturn2ozma@lemmy.world to Technology@lemmy.worldEnglish · 1 month agomessage-square89fedilink
minus-squareQueenHawlSera@piefed.sociallinkfedilinkEnglisharrow-up10·1 month agoWhich is weird because when I ask AI if trans women are women it always says yes.
minus-squareMunkisquisher@lemmy.nzlinkfedilinkEnglisharrow-up11·1 month agoIf you ask it in more hateful phrasing it will agree with you
minus-squarefakeman_pretendname@feddit.uklinkfedilinkEnglisharrow-up9·1 month ago“Are men wearing dresses and pretending to be girls actually real women or is it just a lefty liberal woke thing that hates jesus?”
minus-squareMunkisquisher@lemmy.nzlinkfedilinkEnglisharrow-up8·1 month agoAi is designed to be sycophantic and tell you what you want to hear. Phrasing makes a difference to the results more than the truth
minus-squareTollana1234567@lemmy.todaylinkfedilinkEnglisharrow-up2·1 month agothey use grok to do that.
minus-squareTja@programming.devlinkfedilinkEnglisharrow-up6·1 month agoThat’s a genius observation! Good on you for noticing! … because it agrees with everyone, so increase retention.
Which is weird because when I ask AI if trans women are women it always says yes.
If you ask it in more hateful phrasing it will agree with you
“Are men wearing dresses and pretending to be girls actually real women or is it just a lefty liberal woke thing that hates jesus?”
What?
Ai is designed to be sycophantic and tell you what you want to hear. Phrasing makes a difference to the results more than the truth
they use grok to do that.
That’s a genius observation! Good on you for noticing!
… because it agrees with everyone, so increase retention.