“Meta has been a perennial home for Russian, Chinese and Iranian disinformation,” claims Gordon Kravitz, co-CEO of NewsGuard, a company that provides a tool for assessing the credibility of online information. “Now, Metta has clearly decided to open the floodgates fully.”
Again, fact-checking is not perfect; Krowicz says NewsGuard has already tracked several “false stories” on Meta’s platforms. And the community notes model with which Meta will replace its fact-checking battalion could still be somewhat effective. but research It has been shown by Mahavedan et al. that crowding solutions miss many false positives. And unless Meta commits to greater transparency in how its version is implemented and used, it will be impossible to know whether the systems are working at all.
It’s also unlikely that switching to community notes will solve the “bias” problem Meta executives are so outwardly concerned about, since it seems unlikely to exist in the first place.
David Rand, a behavioral scientist at MIT, said, “Metta’s policy changes and Musk’s motivation to take over Twitter are accusations that social media companies are biased against conservatives.” “There is no good evidence for that.”
In a recently published the paper In Nature, Rand and her co-authors found that while Twitter users who used Trump-related hashtags in 2020 were four times more likely to eventually be suspended than those who used pro-Biden hashtags, they The probability of being shared was also very high. low-quality” or misleading news.
Rand says, “Just because someone is being acted upon doesn’t mean there is bias. “Crowd ratings can do a pretty good job of reproducing fact-checking ratings. . . . You’re still going to see more conservatives get approved than liberals.”
And while X gets some attention thanks to Musk, remember that it’s an order of magnitude smaller than Facebook’s 3 billion monthly active users, which will present its own challenges when Meta implements its own community note-style system. will establish “There’s a reason there’s only one Wikipedia in the world,” Metzarlis says. “It’s very difficult to get anything crowdsourced off the ground at scale.”
As for loosening Meta’s hateful conduct policy, that is itself an internal political choice. It is still allowing some things and disallowing others; Moving those boundaries to accommodate bigotry doesn’t mean they don’t exist. This means that the meta is more accurate than it was back in the day.
A lot depends on how the meta’s system will work in practice. But amid changes in moderation and changing community guidelines, Facebook, Instagram, and Threads are leading to a world where one can say that gay and trans people have “mental illness“Where the AI slope will spread ever more aggressively, where outrageous claims spread without scrutiny, where the truth itself is vulnerable.
You know: like X.
