Meet the Next Fact-Checker, Debunker and Moderator: You

 

Meet the Next Fact-Checker, Debunker and Moderator: You

 

Meta is joining X and YouTube in shifting moderation to users. Are you ready?

 

By Stuart A. Thompson and Kate Conger

Stuart Thompson covers the spread of false and misleading information online from New York. Kate Conger covers X from San Francisco.

Jan. 7, 2025

https://www.nytimes.com/2025/01/07/technology/meta-facebook-content-moderation.html

 

Meta would like to introduce its next fact-checker — the one who will spot falsehoods, pen convincing corrections and warn others about misleading content.

 

It’s you.

 

Mark Zuckerberg, Meta’s chief executive, announced Tuesday that he was ending much of the company’s moderation efforts, like third-party fact-checking and content restrictions. Instead, he said, the company will turn over fact-checking duties to everyday users under a model called Community Notes, which was popularized by X and lets users leave a fact-check or correction on a social media post.

 

The announcement signals the end of an era in content moderation and an embrace of looser guidelines that even Mr. Zuckerberg acknowledged would increase the amount of false and misleading content on the world’s largest social network.

 

What is a Community Note?

On X, the notes are appended beneath posts to add missing details or corrections. This one, highlighted in red, appeared on a post by Elon Musk.

 

“I think it’s going to be a spectacular failure,” said Alex Mahadevan, the director of a media literacy program at the Poynter Institute called MediaWise, who has studied Community Notes on X. “The platform now has no responsibility for really anything that’s said. They can offload responsibility onto the users themselves.”

 

Such a turn would have been unimaginable after the presidential elections in 2016 or even 2020, when social media companies saw themselves as reluctant warriors on the front lines of a misinformation war. Widespread falsehoods during the 2016 presidential election triggered public backlash and internal debate at social media companies over their role in spreading so-called fake news.

 

The companies responded by pouring millions into content moderation efforts, paying third-party fact-checkers, creating complex algorithms to restrict toxic content and releasing a flurry of warning labels to slow the spread of falsehoods — moves seen as necessary to restore public trust.

 

The efforts worked, to a point — fact-checker labels were effective at reducing belief in falsehoods, researchers found, though they were less effective on conservative Americans. But the efforts also made the platforms — and Mr. Zuckerberg in particular — political targets of President-elect Donald J. Trump and his allies, who said content moderation was nothing short of censorship.

 

Now, the political environment has changed. With Mr. Trump set to take control of the White House and regulatory bodies that oversee Meta, Mr. Zuckerberg has pivoted to repairing his relationship with Mr. Trump, dining at Mar-a-Lago, adding a Trump ally to Meta’s board of directors and donating $1 million to Mr. Trump’s inauguration fund.

 

“The recent elections also feel like a cultural tipping point towards once again prioritizing speech,” Mr. Zuckerberg said in a video announcing the moderation changes.

 

Mr. Zuckerberg’s bet on using Community Notes to replace professional fact-checkers was inspired by a similar experiment at X that allowed Elon Musk, its billionaire owner, to outsource the company’s fact-checking to users.

 

X now asks everyday users to spot falsehoods and write corrections or add extra information to social media posts. The exact details of Meta’s program are not known, but on X, the notes are at first visible only to users who register for the Community Notes program. Once a note receives enough votes deeming it valuable, it is appended to the social media post for everyone to see.

 

“A social media platform’s dream is completely automated moderation that they, one, don’t have to take responsibility for and, two, don’t have to pay anyone for,” said Mr. Mahadevan, the director of MediaWise. “So Community Notes is the absolute dream of these people — they’ve basically tried to engineer a system that would automate fact-checking.”

 

Mr. Musk, another Trump ally, was an early champion for Community Notes. He quickly elevated the program after firing most of the company’s trust and safety team.

 

Studies have shown Community Notes works at dispelling some viral falsehoods. The approach works best for topics on which there is broad consensus, researchers have found, such as misinformation about Covid vaccines.

 

In that case, the notes “emerged as an innovative solution, pushing back with accurate and credible health information,” said John W. Ayers, the vice chief of innovation in the division of infectious disease and global public health at the University of California, San Diego, School of Medicine, who wrote a report in April on the topic.

 

But users with differing political viewpoints have to agree on a fact-check before it is publicly appended to a post, which means that misleading posts about politically divisive subjects often go unchecked. MediaWise found that fewer than 10 percent of Community Notes drafted by users ended up being published on offending posts. The numbers are even lower for sensitive topics like immigration and abortion.

 

Researchers found that the majority of posts on X receive most of their traffic within the first few hours, but it can take days for a Community Note to be approved so that everyone can see it.

 

Since its debut in 2021, the program has sparked interest from other platforms. YouTube announced last year that it was starting a pilot project allowing users to submit notes to appear below misleading videos. The helpfulness of those fact-checks is still assessed by third-party evaluators, YouTube said in a blog post.

 

Meta’s existing content moderation tools have seemed overwhelmed by the deluge of falsehoods and misleading content, but the interventions were seen by researchers as fairly effective. A study published last year in the journal Nature Human Behavior showed that warning labels, like those used by Facebook to caution users about false information, reduced belief in falsehoods by 28 percent and reduced how often the content was shared by 25 percent. Researchers found that right-wing users were far more distrustful of fact-checks, but that the interventions were still effective at reducing their belief in false content.

 

“All of the research shows that the more speed bumps, essentially, the more friction there is on a platform, the less spreading you have of low-quality information,” said Claire Wardle, an associate professor of communication at Cornell University.

 

Researchers believe that community fact-checking is effective when paired with in-house content moderation efforts. But Meta’s hands-off approach could prove risky.

 

“The community-based approach is one piece of the puzzle,” said Valerie Wirtschafter, a fellow at the Brookings Institution who has studied Community Notes. “But it can’t be the only thing, and it certainly can’t be just rolled out as like an untailored, whole-cloth solution.”

 

Dec. 12, 2024

Stuart A. Thompson writes about how false and misleading information spreads online and how it affects people around the world. He focuses on misinformation, disinformation and other misleading content. More about Stuart A. Thompson

 

Kate Conger is a technology reporter based in San Francisco. She can be reached at kate.conger@nytimes.com. More about Kate Conger