General_Effort@lemmy.world to Technology@lemmy.worldEnglish · 9 days ago25 arrested in global hit against AI-generated child sexual abuse materialwww.europol.europa.euexternal-linkmessage-square92fedilinkarrow-up1267arrow-down12
arrow-up1265arrow-down1external-link25 arrested in global hit against AI-generated child sexual abuse materialwww.europol.europa.euGeneral_Effort@lemmy.world to Technology@lemmy.worldEnglish · 9 days agomessage-square92fedilink
minus-squareJuxtaposedJaguar@lemmy.mllinkfedilinkEnglisharrow-up2·8 days agoProbably not, but that’s irrelevant. The point is that no one needs to harm a child to find out if the output is sufficiently arousing.
minus-squareDoPeopleLookHere@sh.itjust.workslinkfedilinkEnglisharrow-up1arrow-down1·8 days agoBut how does it get more authentic without actual input if what’s accurate. It’s not enough to tell and AI that’s somethings wrong. You have to also tell it what was right.
minus-squareJuxtaposedJaguar@lemmy.mllinkfedilinkEnglisharrow-up2·8 days agoIt doesn’t need to get more authentic, it just needs to get more arousing, and we have a perfectly ethical way to measure that. You tell the AI it was “right” if the pedos you show it to get aroused.
Probably not, but that’s irrelevant. The point is that no one needs to harm a child to find out if the output is sufficiently arousing.
But how does it get more authentic without actual input if what’s accurate.
It’s not enough to tell and AI that’s somethings wrong. You have to also tell it what was right.
It doesn’t need to get more authentic, it just needs to get more arousing, and we have a perfectly ethical way to measure that. You tell the AI it was “right” if the pedos you show it to get aroused.