1984@lemmy.today to Technology@lemmy.ml · 1 年前OpenAI proposes a new way to use GPT-4 for content moderation | TechCrunchtechcrunch.comexternal-linkmessage-square11fedilinkarrow-up139arrow-down17
arrow-up132arrow-down1external-linkOpenAI proposes a new way to use GPT-4 for content moderation | TechCrunchtechcrunch.com1984@lemmy.today to Technology@lemmy.ml · 1 年前message-square11fedilink
minus-squareFaceDeer@kbin.sociallinkfedilinkarrow-up5arrow-down3·1 年前IMO, It’s sometimes maybe better to have an AI with consistent principles that it applies universally than a capricious human moderator.
minus-squareFaceDeer@kbin.sociallinkfedilinkarrow-up2arrow-down4·1 年前As long as the biases are explicit and consistent it’s still an improvement IMO.
minus-squaretraveler@lemdro.idlinkfedilinkarrow-up2arrow-down7·1 年前AI will follow the rules blindly without much bias…
IMO, It’s sometimes maybe better to have an AI with consistent principles that it applies universally than a capricious human moderator.
Who trains ChatGPT biases? Humans.
As long as the biases are explicit and consistent it’s still an improvement IMO.
AI will follow the rules blindly without much bias…