Hello again. Another week of tough news. At least you’ve got Plaintext, and at least I’ve got you readers. Let’s stick together.
For now, this weekly column is free for everyone to access. Soon, only WIRED subscribers will get Plaintext as a newsletter. You’ll get to keep reading it in your inbox by subscribing to WIRED (discounted 50%), and in the process getting all our amazing tech coverage in print and online.
The Plain View
On May 4, a group of disaffected Republicans known as the Lincoln Project posted an ad on Twitter, YouTube, and Facebook. Inspired by Ronald Reagan’s classic 1984 “Morning in America” ad, the Lincoln Project’s “Mourning in America” recited a litany of grim statistics with depressing images of pandemic America, laying the blame on President Trump. The president was not happy, attacking the ad that very evening.
A day later, Facebook labeled the ad “partly false,” rejected it as inappropriate, and dramatically depressed its circulation when users tried to share the video for free.
If you have been following Mark Zuckerberg’s statements on political advertising, this might seem puzzling. Despite criticism, he has articulated a public policy of not filtering or even fact-checking political advertisements on the platform. It’s up to users to decide the truth for themselves. “I don’t think that a private company should be censoring politicians or news,” he told Gayle King on CBS.
So why did Facebook refuse to run “Mourning in America” as an ad, and bury it otherwise?
The reason, explains Facebook spokesperson Andrew Stone, is that the Lincoln Project is not an ad from a campaigning politician, but an outside organization. If candidates for public office pay Facebook to circulate even demonstrably false claims, Facebook will happily place it in the News Feeds of a targeted audience. But if the advertiser is not running for office, Facebook will append a scarlet letter to ads identified as making exaggerated claims and misstatements.
But wait: The “Mourning” ad seems accurate. Online critics wondered whether Facebook—whose handling of misinformation in the 2016 election seemed to benefit the Trump campaign—was doing the White House a favor in censoring the Lincoln ad.
The truth is not so nefarious but not terribly comforting, either.
Facebook relies on outside fact-checking organizations to determine the truthfulness of controversial content. These operations choose what stories to vet, either by identifying controversial content or by selecting from a dashboard of popular content provided by Facebook. In this case, Politifact, the fact-checking branch of the nonprofit Poynter Institute, decided to look at the ad, which had instantly garnered a lot of attention.
According to Aaron Sharockman, Politifact’s executive director, his fact-checker didn’t have a problem with any of the many statistics in the ad about the coronavirus death toll or unemployment numbers. Instead, Politifact zeroed in on one sentence: “Donald Trump bailed out Wall Street but not Main Street.” To many (like me) this may seem like an opinion, whose worth depends on data. The Lincoln Project provided a number of sources, including Bloomberg, NBC, Vanity Fair, and even the New York Post, where a Fox Business News reporter wrote in an op-ed, “Wall Street traders will make money, while Main Street businesses face economic conditions not seen since the Great Depression.” But Politifact chose an absolutist interpretation. Because the Cares Act passed by Congress and signed by Trump did some things for Main Street, it reasoned, in effect Trump had indeed bailed out mainstream America. “Most people who argue seem to suggest that maybe Trump has bailed out Wall Street more than Main Street,” says Sharockman. “But that’s not what the ad said. So I feel real good about the rating as calling it false.”
On Facebook the warning label read “Partly False.” I asked Sharockman, since every other sentence was indisputably factual, why the ad wasn’t labeled “Mostly True.” He told me that the only alternatives were “True, Partly False, and False.” (Later I did my own fact-check: Politifact’s website describes a “Truth-o-Meter” that includes categories like “Mostly True” or “Half True.” Apparently, Facebook accepts only the three that Sharockman mentions.)
Fact-checking of political ads can be more art than science; it often rests on slender distinctions. But as Politifact knows, the power of those distinctions can become grotesquely distorted when they are translated to labels that Facebook blindly applies. The penalties are severe. When an ad is deemed False or Partly False by a fact-checking organization, Facebook will pull the advertisement. Even worse, when people share the ad with friends, it is treated as toxic content and buried in the News Feed. When people do see the post, they must click through the warning label to view the actual ad, as if it contained gory medical scenes or other disturbing content.
The Lincoln Project complained to Facebook and got no formal response beyond a referral to the fact-checkers. Politifact’s stance is that the organization should have changed the ad to say that Trump helped Wall Street more than Main Street. Sharockman says the fix would take only six seconds. Jennifer Horn, a cofounder of the Lincoln Project, says that her organization will not bow to censorship. (There is also the fact that this controversy has made the ad even more effective—the Lincoln Project says that it has been a champion in eliciting donations.) She notes that YouTube and Twitter have not objected to the content of the ad, and several television markets have run it without asking for changes.
Only Facebook has effectively banned it. Ironically, If Joe Biden had placed an identical ad, Facebook would not have impeded it or given it a warning label. Even if Biden said that Trump personally breathed deadly germs on 60,000 Americans, Facebook would have let that stand. Complete political freedom!
Or, depending on who is speaking, censorship by nitpick.
In 2007, Facebook announced its first significant ad strategy. I wrote about it in Newsweek. Incidentally, what I did not know at the time was that the internal codename for the ad products announced in the fall of 2007 was “Pandemic”:
With its new program, Facebook announced that it was empowering advertisers to target those ads using the information on the personal profiles that members supply to Facebook. A national advertiser could sell ads to a huge group (all women between 25 and 40), or a local advertiser, like a restaurant, could pay much less to reach a microgroup (Ivy League-educated Indian-food lovers in a specific ZIP code). You could even target people who work for a specific company; Facebook itself has used this feature to solicit employees from its competitors. It’s an innovative strategy. But since Facebook users originally supply that information to share with friends, and not with advertisers, they may believe that utilizing those details for ad targeting wasn’t part of the deal. (Facebook’s privacy officer, Chris Kelly, says that the personal data itself is not given to the advertisers.)
Ask Me One Thing
Max asks, “Hey Steven! I saw your Plaintext posted on the site today, and I was trying to figure out how I actually subscribe to get it sent to me. I’ve been a subscriber of WIRED for a few months now, and I’m sad to say I wasn’t subscribed to Plaintext, so how do I do that?”
Max, WIRED subscribers should get Plaintext automatically—unless they opted out of getting mail from us, which is their absolute right. Maybe when you opted out you didn’t realize that you would be turning down Plaintext in your inbox every week. In any case, we hope to build an option that lets opted-out subscribers sign up for this newsletter. Can’t say when, but it’s on the list. And as a bonus for pointing this out to us, we’ve managed to get you signed up right now! Just don’t hit “Unsubscribe” by mistake.
You can submit questions to firstname.lastname@example.org. Write ASK LEVY in the subject line.
Ever since the Supreme Court began conducting oral arguments by phone, Justice Clarence Thomas has been asking more questions than he has for over a decade. That would mildly qualify as a sign of apocalypse on its own, but this week he invoked the name Frodo Baggins as the choice of a hypothetical electoral college voter. End-Times gold!
Last but Not Least
Andy Greenberg’s epic story of a hacker who saved the internet and then got busted by the FBI is worth a good chunk of your lockdown time.