The hard and true limits of checking facts

Third-party fact-checking on social media was supposed to be a silver bullet. It’s not turning out that way.

Hey, there. Thanks for having me in you inbox. This is Slow Build, a newsletter on tech, society, and a splash of design by Nancy Scola. A programming note: we’re taking this thing biweekly — on Wednesday and Fridays. If this was forwarded to you or you’re reading online, now’s the time to go ahead and subscribe

So let’s talk about whether fact-checking can redeem social media, as the idea’s having a mini-revival this week. 

Just a bit of context to start. For years, the notion’s circulated that one way improving the quality of information circulating on platforms like Facebook and Twitter is to inject third-party fact checkers into the mix. Those fact checkers are sometimes journalists; often they work for shops that exist solely to check facts, like PolitiFact or They assess whether what’s floating around online is true, and once those decisions have been rendered, the tech companies sometimes operationalize it. The dynamic of how that works is constantly shifting, but it can mean that the social platforms append little notes telling users that the post they’re seeing is disputed. It can also mean that the platforms somehow throttle the spread of the post, or limit the way users can engage with it. 

This isn’t happening just in the United States. Over on Rest of World, the journalist Sonia Faleiro had a robust piece this week on how fact checkers in India are struggling to gain traction. She details in particular an incident where a politician in Prime Minister Narendra Modi’s Bharatiya Janata Party shared a video of a vicious street attack, suggesting the perpetrator was a Muslim man living in India. A fact-checking outfit called Alt News quickly found that the clip instead showed a person with mental illness lashing out in Sri Lanka. Alt News posted its findings. Nothing happened. 

(If you’re not familiar with Rest of World, it’s a year-old non-profit news site that takes its name tongue-in-cheekily from business-speak that distinguishes between the U.S., Europe, and, you know, everybody else.)

Back in the U.S., also this week a Michigan Republican state rep introduced a novel bill, called the “Fact Checker Registration Act” that would, well, require just that — as well as require members of the International Fact Check Network to file a million-dollar bond against which those legally harmed by a fact check could, somehow, legally draw. 

And members of that International Fact Check Network took part this week in an event called “United Facts of America: A Festival of Fact-Checking.” At it, CNN’s chief media correspondent Brian Stelter said, “I think (fact-checkers) ran the equivalent of multiple marathons in the past few years. I think that those marathons have won over a lot of supporters and people are cheering on this movement and wanting more of it.”

That’s an eye-catching quote because if you’re talking about journalists doing more robust veracity-checking in the course of their work, that perhaps ring true. But if you’re talking about independent fact checkers, I’m not so sure. And after years of reporting on these folks, and their interplay with the big social media platforms, I’m not so sure the fact checkers are such about it either. 

It’s an act of desperation as much as anything else, and I think there are a number of under-explored reasons why.

Judgement calls: “Lyon is the capital of France” is an easy statement to fact check. But problem online isn’t with people misidentifying the capital of France. Even determining whether or not a video was shot in India is a straightforward determination of truth. And while the Trump era was something of halcyon days, because obvious and egregious falsehoods were everywhere, we’re probably returning to the norm of ten years ago where Democrats and Republicans shared the belief that fact-checkers routinely made wrong calls about their nuanced arguments. 

Take a look, for example, at the Washington Post’s Fact Checkers’ analysis of President Biden’s recent joint address to Congress. Did Biden have his facts wrong? Eh, kinda? Sometimes, sorta, depending on how you look at it… One of the fact checkers’ objections has to do with whether presidents of the United States can claim credit for job growth. We’re not in “the sky is green” territory here. And boiling down these kinds of assessments to where they’re usable in the social-media context is tough. 

Cultural context: The U.S. is often treated as the test bed for debates about online content, and other places are properly thought of as sort of marginal cases. On numbers alone, that’s seriously flawed. More people use Facebook in India than in the U.S. As an old anthropology major, I’m used to the complaint that cultural relativism is a dodge. But it really is enormously difficult to make reasoned, granular assessments about cultures other than your own.

One thing to look at here is the situation with the New York Post’s story on Hunter Biden just before the 2020 election. Let me tell you, oy, what a mess. Both Twitter and Facebook quickly made decisions about the validity of the information and the dangers of it spreading, with both restricting users ability to post it for a time. Just about no one thinks it was handled well. 

And we probably should ask ourselves whether California companies that couldn’t manage a story in a high-profile New York City-based publication about a major party candidate’s well-known son at the height of a U.S. election are well-situated to make sense of things happening on the ground in, say, Thailand. If the solution is to push that decision-making local, it’s not at all clear that the companies are capable of the necessary oversight. 

Scale: Frankly, it’s long been a mystery to me why the tech companies don’t point out that holding them responsible for the literally billions of pieces of information posted to their networks each week is a very tall order — and make the argument that if users want to be able to instantaneously publish their thoughts for all the world to see, the trade-off is it’s sometimes going to be a bit of a mess. 

In that vein, one powerful takeaway from my reporting on the rise of fact checkers in recent years is that these organizing are small. We’re talking a handful of people here and there. The Rest of World’s Faleiro, in her piece on India, touches on this, writing that “tech companies worth many billions of dollars had in effect outsourced their responsibility to a tiny group of people.” 

In the U.S., at least, some of these groups have gotten funding from the tech companies to help stay afloat. That itself is a challenge, even putting aside any conflicts of interest. If there’s not enough public or philanthropic demand to support these institutions, their reach is always going to be limited. 

Impact: Social media is often a machine for confirming our biases and self-conceptions. We all do it, seeking out content that affirms or amplifies our existing sense of self. And there’s little evidence so far that fact-checking has any real ability to make us rethink what we already ‘know.’

Exacerbating that is that politicians around the world are realizing that spreading a warped version of reality online can change the world in their favor. Of course, propaganda is hardly new. But the modern political leader is figuring out exactly how to weaponize information for a national, even global audience that has proven it isn’t much invested in whether it’s true. Facebook and Twitter made a big deal about stating conclusively who won the 2020 U.S. presidential election. Let’s just say that didn’t stick. 

“Misinformation is a challenge globally, but in India, it’s practically baked into the ruling party’s communications,” writes Faleiro. And when it comes Alt News, “they still find themselves playing whack-a-mole in a country increasingly hostile to the truth.”

Accountability theater?: For social media companies, the appeal of third-party fact-checking was obvious enough. They wouldn’t have to act as judge and jury. They were just implementing the decisions made by experts. And even if the practice hasn’t been the silver bullet they perhaps once wished it would be, it’s easy to argue that it’s helped check the distribution of the worst of the worst.

But there’s one potentially worrisome possibility that’s beginning to emerge: that the outsourcing of fact-checking gives tech companies a veneer of accountability that takes pressure off them to responsibly operate their networks as a whole. We might include Facebook’s Oversight Board in that ‘accountability theater’ bucket, too. One more and we have a trend.


With fact-checking sucking up a lot of oxygen, it’s probably worth considering whether it’s worth it. But it’s easy to highlight where it isn’t working than it is to point to alternatives.

Here’s a thought. I’ve been spending time lately on LinkedIn after, frankly, spending many years actively ignoring it. As much as the platform can be a carnival of self-promotion, I’m struck by how it’s truly about human connection. It’s fascinating to know what people you know and kinda know are up to, how you might plug into it, and how it might inform your own work. LinkedIn’s legitimately about engaging in the experiences of other people, and you’re unlikely to go on there to spread utter nonsense.

A big part of that is design choice. For LinkedIn, I think it matters a great deal that you’re always one click away from your work history and present. Twitter and Facebook have started experimenting with design adjustments meant to similarly boost users’ sense of connection what they post, like the read-before-resharing prompt that Twitter’s had for a while and Facebook introduced this week

Those sort of tweaks to introduce a bit of friction here and there potentially add up to higher-quality information across these networks. But they’re arguably a greater investment of time and resources than simply appending notices saying, “Someone else says this thing right here is wrong.”

Reading promiscuously:

—Robinhood’s origin story is well-told in this deep dive by the New Yorker’s Sheelah Kolhatkar. The author digs into the thinking behind the trading app’s founders’ now controversial decision to pay for itself through “payment for order flow.” (PFOF, Kolhatkar usefully explains, “is pronounced ‘P-fof.’”) The piece gets at the fascinating debate over what, exactly, Robinhood is up to. Here’s Benchmark Capital’s Bill Gurley: “My issue with Robinhood is, I think their mission and what they say they stand for is not actually true.” And here’s Howard Lindzon of Social Leverage, a Robinhood investor, making the Uber-and-the-taxi-industry argument: “It’s because of the mistakes the incumbents made that Robinhood even exists.” 

One dominant thread in the piece is how Robinhood is managing its new responsibilities, including whether having better customer service would help it avoid some of the troubles that have affected it thus far. A Slow Build prediction: like where tech companies once put off forming policy teams until the very last possible moment but now include them in their launch plans, having even a semblance of powered-by-humans customer service will increasingly become an expectation for even the smallest of startups. 

New York’s Benjamin Wallace peels back the curtain on the love affair between the tech industry and journalists, and Silicon Valley’s growing belief that there’s little downside to messing with legacy media. “And so a war is on,” writes Wallace, “between the tech titans and a relentless generation of largely digital-native reporters looking to speak truth to power while racking up Twitter followers in the process.”

—As the result of an obscure provision in the stimulus passed by Congress, local governments now can get .gov websites for free. Until now, they had to pay $400 a year, which some jurisdictions found too big an obstacle. Exciting stuff, I know! 

But with the country’s critical infrastructure vulnerable — and thus four hundred bucks getting you about two tanks of gas at the moment — the move could both boost the security of government digital resources as well as the public’s faith in the information and services we get from local government. Tiny bit of legislation, big impact. 

See you next week, and thanks for reading.