EatingOnions@lemmy.world to World News@lemmy.worldEnglish · 15 天前Companies Are Using Reddit to Manipulate ChatGPT and Google AI Searchwww.404media.coexternal-linkmessage-square28linkfedilinkarrow-up1244arrow-down10 cross-posted to: technology@lemmy.worldreddit@lemmy.world
arrow-up1244arrow-down1external-linkCompanies Are Using Reddit to Manipulate ChatGPT and Google AI Searchwww.404media.coEatingOnions@lemmy.world to World News@lemmy.worldEnglish · 15 天前message-square28linkfedilink cross-posted to: technology@lemmy.worldreddit@lemmy.world
minus-squareCovertOperative@piefed.ziplinkfedilinkEnglisharrow-up5arrow-down1·15 天前Ok, what? How is increasing the amount of wrong information on the web not the opposite of helping?
minus-squaredaannii@lemmy.worldlinkfedilinkEnglisharrow-up6arrow-down1·14 天前It’s called poisoning the data
minus-squarebandanawearingbanana@lemmy.worldlinkfedilinkEnglisharrow-up2·edit-29 天前But most of this data has already been scraped by LLMs. It is stupid and petty to edit one comment which mostly real people looking for a spefic answer to a tech problem will see.
minus-squaredaannii@lemmy.worldlinkfedilinkEnglisharrow-up1arrow-down1·9 天前Data scrapping happens continuously And poisoning is very effective. Especially if more people do it.
Ok, what? How is increasing the amount of wrong information on the web not the opposite of helping?
It’s called poisoning the data
But most of this data has already been scraped by LLMs. It is stupid and petty to edit one comment which mostly real people looking for a spefic answer to a tech problem will see.
Data scrapping happens continuously And poisoning is very effective. Especially if more people do it.