Since Donald Trump regained the presidency on Nov. 5, a parade of Silicon Valley luminaries has been engaging in an unseemly humiliation fest, making pilgrimages to Mar-a-Lago, contributing millions to his inaugural fund and meddling in editorial departments. of the publications they have in an apparent attempt to curry favor with the new leader. Yesterday, the turn was for the founder and CEO of Meta, Mark Zuckerberg.
In a five-minute Instagram video, sporting his new curly hairstyle and a $900,000 Gruebal Forsey watch, Zuckerberg announced a series of dramatic policy changes that could open the floodgates of misinformation and hate speech on Facebook, Threads and Instagram. His reasoning parroted arguments that right-wing lawmakers, pundits and Trump himself have been repeating for years. And Zuckerberg wasn’t shy about the moment, explicitly saying that the new political regime was a factor in his thinking: “The recent election also feels like a cultural turning point toward once again prioritizing expression,” he noted in the video.
According to Zuckerberg, the main impetus for the change is the desire to encourage “freedom of expression.” “Meta’s social networks had become too extreme in restricting users’ expression, he said, so the idea Central to the changes – which included ending Meta’s multi-year partnerships with third-party fact-checking organizations and withdrawing from efforts to slow the spread of hate speech – is letting freedom ring, even if it means “let’s catch less things “bad.”
But the key is in Zuckerberg’s nomenclature. He called his company’s (not entirely successful) efforts to prevent the promotion of toxic content “censorship,” and has now adopted the same bad-faith characterizations of his employees’ work as the political right, which used him as a cudgel. to force Facebook to allow ultraconservatives to promote things like targeted harassment and intentional misinformation. In reality, Meta has every right to police its content however it wants: “censorship” is something governments do, and private companies simply exercise their own free speech rights by deciding what content is appropriate for their users and advertisers.
Zuckerberg indicated for the first time that he might agree to the term in a nice letter who wrote last August to Republican Congressman Jim Jordan, in which he said that the Biden administration wanted Meta to “censor” some content related to the covid-19 pandemic. (The content remained, which actually illustrates that Facebook has the power to shape free speech in the US, not the government.) However, in his Instagram post yesterday, Zuckerberg used the term synonymously with the entire practice of content moderation: “We are going to dramatically reduce censorship on our platforms,” he promised. An alternative reading could be: let’s let the doberman!
From center to right
In the same letter to Jordan, the former left-wing CEO vowed that he would not align himself with any political party: “My goal is to be neutral and not play any role, nor even appear to play any role,” he wrote. Now that Trump has been elected, all that has gone away: “It feels like we’re in a new era,” he said in yesterday’s video. Apparently, it is an era where private companies change their rules to ensure they are in line with the party in power. In the last week alone, Zuckerberg replaced Nick Clegg, the company’s former president of global affairs, with Joel Kaplan, a former member of the Republican Party and clerk to the late Justice Anthony Scalia, who in his day he urged Facebook to ignore misinformation during the 2016 election. Zuckerberg also named the president of the Ultimate Fighting Championship, Dana Whitean ardent Trump supporter, to serve on Meta’s board.
Another indication that there is a MAGA element to these changes is Zuckerberg’s announcement that he is moving Meta’s trust and safety and content moderation teams from California to Texas. Again, he said out loud that the reasons for the geographic move were political: “I think it will help us build confidence to do this work in places where there is less concern about the bias of our teams.” Hello, Mark? This move simply anchors Meta’s content referees in a potentially biased place. different. It’s also a striking statement that Zuckerberg himself might consider California, Trump’s kryptonite, a less tasty place to work than deep red Texas.
Ironically, all of this is happening at a time when Meta has managed to improve its previously dismal record of removing hate speech and other legal but toxic posts. In 2019, when I spent a day in one of the company’s content moderation offices, the beleaguered moderators working there told me that dealing with gray areas like hate speech was the one thing AI could never do. But more recently, Meta’s chief AI scientist, Yann LeCun, revealed that generative AI tools had changed the game: “Five years ago, of all the hate speech that Facebook removed from the platform, between 20 and 25 percent were removed preventively by AI systems before anyone saw it,” he told me in an interview in late 2023. “Last year, it was 95%.”
Lax moderation
Now, Meta declares that he is throwing that victory in the trash. The algorithm will change, and the AI will no longer proactively block “less serious” policy violations, such as hate speech and, presumably, harassment. (Meta says it remains committed to vigilant blocking of “high severity” illegal content, such as terrorism and child exploitation.) Meta has also made a number of changes to its speech policies that allow certain types of comments that were previously considered hateful. One of them affects a hot topic for the right: gender and sexual identity. For example, it is now okay to claim that homosexuals or transsexuals are mentally ill or simply call them “weird” on Facebook.
To be fair, some elements of what Zuck is proposing make sense, but he is mischaracterizing the past. It is true that his fact-checking experiment has not been entirely successful, but one of the reasons for the failure was Meta’s inability to fully engage, in part due to criticism from the right. There have been cases where overzealous algorithmic zeal has stifled useful debate, as well as many cases where permitted expression has been wrongly removed. But Meta’s silly blocking of talking points like “women should not serve in combat roles” (which has now been reversed) was typically less the result of flawed rules or algorithms than of insufficient investment in human content moderation, that can make more nuanced distinctions.
#factchecking #hate #speech #Meta #MAGA
