Fact-checking has become partisan. Can it survive the backlash from conservatives and Big Tech?

In a coffee-table book published last year about his first term in office, U.S. president-elect Donald Trump threatened to jail Mark Zuckerberg, suggesting the Meta CEO had helped rig the 2020 election.

The conspiracy theory had circulated widely on social media, including on Meta’s own platforms, Facebook and Instagram. It was eventually debunked by one of the third-party groups that Meta paid to fact-check popular content on its sites.

On Tuesday, Zuckerberg announced an abrupt end to Meta’s fact-checking program in the U.S., drawing praise from Trump.

Zuckerberg’s move appeared aimed, in part, at shielding Meta from an escalating effort by Republican lawmakers and activists to cripple the fact-checking industry that has arisen alongside social media.

It’s also causing a reckoning among fact-checkers themselves about the value and effectiveness of their work amid the daily tidal wave of falsehoods.

“Fact-checking has been under attack. It’s been made into a bad word by some corners of our politics in the U.S. and around the world,” said Katie Sanders, editor-in-chief of PolitiFact, which until this week had been one of the partners in Meta’s fact-checking program. 

“We’re still in the very earliest stages of untangling the implications. But there’s anxiety in the air, for sure.”

‘Let’s just label it’

Fact-checking has been a routine feature in news media since at least the 1930s.

But as social media platforms grew in popularity in the 2000s, there emerged a number of publications — such as FactCheck.org and PolitiFact — dedicated almost entirely to verifying the statements of public figures. 

The election of Donald Trump in 2016, however, proved to be a watershed moment for this emerging industry. 

The candidate’s penchant for uttering falsehoods, alongside concerns about social media being used by foreign actors to manipulate public opinion, generated intense pressure on companies like Facebook to take action. 

Facebook entered into partnership agreements with several fact-checking outlets to review content it flagged as potentially misleading. The program eventually expanded to around 130 other countries, including Canada. 

“People really thought, let’s just label it. We should just tell people what’s false, what’s not, and that’s going to solve the problem,” said Katie Harbath, a former director of public policy at Facebook. 

“But immediately there were challenges with the fact-checking program. They’re not able to do it quickly and they’re not necessarily able to do it at scale.” 

Those shortcomings were often the source of frustration for liberals, who felt too much misinformation was falling through the cracks. Many conservatives, on the other hand, believed their content was being unfairly targeted for verification.

Republican-led backlash

In recent years, suspicion of fact-checking programs has turned to outright hostility.

Congressional Republicans and conservative activists targeted The Election Integrity Partnership, a fact-checking coalition of academics and other experts, with so many legal demands that it effectively stopped operating last June. 

Trump’s pick to lead the Federal Communications Commission, Brendan Carr, has spent several weeks attacking the fact-checking efforts of large tech companies. He accused them of supporting a “censorship cartel” and threatened regulatory action.

Carr singled out NewsGuard, a company that rates the credibility of news sites and has given low scores to pro-Trump outlets that circulated false claims about the 2020 election, such as NewsMax. (Other conservative outlets, including Fox News and the New York Post, are rated as trustworthy.)

“Everybody is harmed by misinformation … whether the misinformation harms the left or harms the right, because it means that people are operating with less full understanding of the underlying facts than they should have,” said NewsGuard co-CEO Gordon Crovitz, a lifelong Republican and former publisher of the Wall Street Journal. 

“I think this is very much a bipartisan issue. It’s taken on something of a partisan hue for the moment in the States, but I think that is fleeting. Trusted information is important for all sides in democracies.” 

Zuckerberg gets fact-checked

Meta’s decision to kill the fact-checking program was part of a broader set of changes aimed at loosening content restrictions in the name of “free speech.”

These included new policies that allow users to call LGBTQ people mentally ill or abnormal.

In the five-minute video announcing the changes, Zuckerberg said Meta’s fact-checkers were “too politically biased.”

Ending the program, he added, will “dramatically reduce the amount of censorship on our platforms.”

His reasoning, not surprisingly, came under scrutiny from fact-checkers. 

They pointed out that the partners in the program never removed content from Meta’s sites. Their work only appeared as a warning attached to content that had undergone a thorough review.

“We have a really rigorous process for testing the claims that we set out to fact-check. We have a plan going in for how we will learn about this topic and get the definitive answer,” said Sanders. “It requires time — and expertise, frankly.”

It was ultimately Meta’s decision whether to remove content or shut down a page, something the company rarely did, according to Sanders.

Much of what fact-checkers flagged on a daily basis wasn’t political speech per se, but rather scams and other forms of clickbait, said Alexios Mantzarlis, director of the Security, Trust and Safety Initiative at Cornell Tech, a research centre in New York.

“That was the kind of stuff that this program was meant to solve. It wasn’t meant to solve political lying, which is as old as humanity,” said Mantzarlis, a former director of the International Fact-Checking Network, which helped Facebook set up its fact-checking program.

PolitiFact’s work for Meta included correcting information about mass shootings, natural disasters and ineffective or dangerous health remedies. 

“I would just expect it to become a junkier environment when those claims can proliferate unchallenged,” Sanders said.

Zuckerberg said the fact-checking program will be replaced by a process similar to Community Notes, the crowd-source approach used on X.

While crowd-sourced fact-checking can be effective with the right incentives, the Community Notes feature on X is mainly a forum for further partisan bickering, Mantzarlis said. 

“The particular irony of Zuckerberg throwing fact-checkers under the bus as ‘partisan’ is that his proposed alternative does not seem like a haven for bipartisanship and Kumbaya getting together,” he said.

With high supply comes high demand

For the moment, Meta is only ending its fact-checking program in the U.S. A division of Agence France-Presse provides fact-checking for Canada and continues to operate. 

“It’s a hard hit for the fact-checking community and journalism. We’re assessing the situation,” AFP said in a statement following Zuckerberg’s announcement.

Meta was a major funder of fact-checking operations in the U.S., and its retreat will likely trigger a reordering within the industry, said Sanders.

“But it’s not something that can be killed. It’s here to stay, regardless of whether people in power like it or don’t,” she said. 

In fact, given the endless supply of misinformation, demand for fact-checking has never been higher from advertisers, said Crovitz.

“There’s a tremendous amount of misinformation, whether it’s from Russia, China, Iran or from generative AI models hallucinating,” he said. 

“And there are an increasing number of entities that are concerned about misinformation and want to be sure they’re not contributing to it.”