Fact-Checking Was Too Good for Facebook
Yesterday, Meta CEO Mark Zuckerberg announced that Facebook would end fact-checking on its platform. In the process, a partnership with the network of third parties that has provided review and ratings of viral misinformation since 2016 will be terminated. To some observers, this news suggested that the company was abandoning the very idea of truth, and opening its gates to lies, perversions, and deception. But this is wrong: Those gates were never really closed.
The idea that something called “fact-checking” could be (or could have been) reasonably applied to social-media posts, in aggregate, is absurd. Social-media posts can be wrong, of course, even dangerously so. And single claims from single posts can sometimes be adjudicated as being true or false. But the formulation of those distinctions and decisions is not fact-checking, per se.
That’s because fact-checking is, specifically, a component part of doing journalism. It is a way of creating knowledge invented by one particular profession. I don’t mean that journalists have any special power to discern the truth of given statements. Naturally, people attempt to validate the facts they see, news-related or otherwise, all the time. But fact-checking, as a professional practice linked to the publication of news stories and nonfiction books, refers to something more—something that no social-media platform would ever try to do.
[Read: This is how much fact-checking is worth to Facebook]
Here at The Atlantic, every story we put out goes through a fact-checking process. That usually takes place after the story has been reported, written, and edited. Some of that process is pretty straightforward: A quote from a source might be verified against an interview recording or transcript; dates, locations, or statistics might be compared to the sources from which they were drawn.
Other aspects of the process are more discursive. Is the writer’s sentence fairly paraphrasing someone’s statement? Does it—and the publication—mean to present that person’s statement as informative, dubious, or something else? Sometimes additional research, follow-up interviews, and internal negotiations will be required. In some cases, fact-checking has more to do with evaluation, judgment, and wordsmithing than getting any single line “right” or “wrong.” The process can be very strange. It’s often time-consuming.
Outside of newsrooms, though, fact-checking has come to have a different meaning, and a smaller scope. It may describe the surface-level checks of claims made by politicians in live debates—or of assertions appearing in a dashed-off post on social media. Small-bore inspections like these can help reduce the spread of certain glaring fabrications, a potential benefit that is now excluded from Meta’s platforms by design. But that’s a whack-a-mole project, not a trust-building exercise that is woven into the conception, research, authorship, and publication of a piece of media.
Fact-checking, in this broader sense, assumes its practitioners’ good-faith effort to find or construct truth, and then to participate in the interactive process of verification. When done seriously and deliberately, it imbues a published work with an ethos of care. Journalists retain detailed records of their reporting, annotate them, and submit them with the stories they file. They may be asked to provide additional support or to consider possible objections. The scope of each claim undergoes consideration. Scene-setting—writing that describes a situation or environment—will be subjected to the fact-check, too. “Even the bathroom wallpaper had a bovine theme,” I wrote about a filling-station bathroom in a profile of the children’s author Sandra Boynton, who puts lots of cows in her books. The fact-checker asked if I could prove it. Having anticipated the question, I had taken a photo in the filling-station restroom. Would we have printed the line had I not done so? That’s not the point. Rather, such evidentiary concern suffused the entire effort, not just the part where someone made sure I wasn’t lying.
This process sometimes fails. It may be foiled by sloppiness or haste. But many posts on social media lack even the aspiration to be true. Some people posting may intend to mislead, coerce, or delude their audiences into believing, buying, or simply clicking. Others are less malicious, but still, as a rule, they are not engaged in journalism and do not necessarily share its values. That makes their content not lesser, but different in kind. On social media, people share their feelings, the things they saw, the images they made of the activities they performed (or pretended to perform). They comment, like, and share posts that spark delight or fear, and they may do so without too much concern for their effects on other people’s choices or opinions.
As I’ve written before, giving everyone with a smartphone the ability to say anything they want, as often as they want, to billions of people, is a terrible idea. In the deluge that results, verification is impossible. Sure, one might take the time to affirm or reject the truth of a tiny subset of the claims posted to a platform, but even modest efforts run afoul of the fact that different people post for different reasons, with different goals.
The effort Facebook attempted under the name fact-checking was doomed. You can’t nitpick every post from every random person, every hobby website, every brand, school, restaurant, militia lunatic, aunt, or dogwalker as if they were all the same. Along the way, Facebook’s effort also tarnished the idea that fact-checking could be something more. The platform’s mass deployment of surface-level checks gave the sense that sorting facts from falsehoods is not a subtle art but a simple and repeating task, one that can be algorithmically applied to any content. The profession of journalism, which has done a terrible job of explaining its work to the public, bears some responsibility for allowing—even encouraging—this false impression to circulate. But Facebook was the king of ersatz checking. Good riddance.