mesa to [email protected]English • 4 months agoGoogle searches for deleting Facebook, Instagram on the rise after Meta ends fact-checking | TechCrunchtechcrunch.comexternal-linkmessage-square154fedilinkarrow-up11.06Kcross-posted to: [email protected]
arrow-up11.06Kexternal-linkGoogle searches for deleting Facebook, Instagram on the rise after Meta ends fact-checking | TechCrunchtechcrunch.commesa to [email protected]English • 4 months agomessage-square154fedilinkcross-posted to: [email protected]
minus-square@[email protected]linkfedilinkEnglish1•3 months agoMeta did not end fact checking. Meta switched to a system of community notes, which is superior at fact checking than a centralized team of fact checkers.
minus-square@[email protected]linkfedilinkEnglish10•3 months agoSo community notes can address more instances of misinformation, that part is true. But if the community provides misinformation as the “note” then it can actually spread and legitimize misinformation. So superior is definitely the wrong word for it. Perhaps more efficient? But also more likely to reinforce echo chambers. Superior would be implementing community notes and then having those checked by centralized fact checkers.
minus-square@[email protected]linkfedilinkEnglish2•3 months agoThats the nifty part, for centralized fact-checkers we just use LLMs!
minus-square@[email protected]linkfedilinkEnglish2•3 months ago“fact”-checkers… This is probably a good prefilter, but definitely one of the few (or rather many) things that LLMs are pretty bad in…
Meta did not end fact checking.
Meta switched to a system of community notes, which is superior at fact checking than a centralized team of fact checkers.
So community notes can address more instances of misinformation, that part is true.
But if the community provides misinformation as the “note” then it can actually spread and legitimize misinformation.
So superior is definitely the wrong word for it. Perhaps more efficient? But also more likely to reinforce echo chambers.
Superior would be implementing community notes and then having those checked by centralized fact checkers.
Thats the nifty part, for centralized fact-checkers we just use LLMs!
“fact”-checkers… This is probably a good prefilter, but definitely one of the few (or rather many) things that LLMs are pretty bad in…