Truth Social Has A Content Moderation Problem

Before 42-year-old Ricky Shiffer made his fatal attempt to breach an FBI office in Cincinnati, an account in his name posted a “call to arms” on the site, Truth Social, and told people to “get whatever you need to be ready for combat.” A letter that the House Committee on Oversight and Reform sent to Truth Social on Aug. 19 expressed concern over other posts, including one that said “F--- the Feds! The Second Amendment is not about shooting deer! Lock and load!” and another calling on people to “Arm yourselves! We are about to enter into Civil War!”

The letter, addressed to former Republican Representative Devin Nunes, the chief executive officer of Trump Media & Technology Group, the parent of Truth Social, demanded information about what criteria the platform uses to decide which content should be removed and the resources it dedicated to content moderation. It also asked how many threats against federal law enforcement it had identified since the Mar-a-Lago search and how many of those posts were removed or reported to authorities. The committee made similar inquiries to Meta Platforms, Twitter, TikTok, and other social media companies.

Founded by Trump after his ouster from Twitter, Facebook, and YouTube in the wake of the Jan. 6, 2021, riot at the U.S. Capitol, Truth Social has promised to be a place where its users can say whatever they please. When it launched in Apple’s App Store in February, the site joined an increasingly crowded field of “alternative tech” platforms catering to right-wing users, including Gettr, Parler, and Gab.

The app has been downloaded about 3.2 million times, trailing far behind Parler and Gettr, according to estimates from Sensor Tower, a firm that tracks app data. (Sensor Tower couldn’t provide estimates for Gab, which isn’t available in the App Store or Google Play Store.) But downloads have increased precipitously since the search of Trump’s home, illustrating how much Truth Social and its peers thrive on political conflict.

There’s been a noticeable intensification in violent rhetoric on all these services since the FBI’s Mar-a-Lago search, says Josephine Lukito, an assistant professor at the University of Texas at Austin. “The FBI raid has increased this need to take action” in users’ minds, she says. “The discourse has ramped up—the level of frustration has ramped up.”

That Truth Social would take a more tolerant view toward speech that mainstream social networks consider beyond the pale is to be expected—that’s essentially why it exists. More surprising is how aggressive it’s become at filtering other kinds of posts. In an Aug. 2 report, Cheyenne Hunt-Majer, a fellow for the consumer advocacy group Public Citizen, wrote that she shared a post on Truth Social containing the phrase “abortion is healthcare,” only to find it was rendered essentially invisible on the site, a process sometimes referred to as shadow-banning. Hunt-Majer wrote that the post reappeared after a TikTok video she made about the incident went viral. But another one of her Truth Social posts—this one reading, “If you don’t own a uterus and know everything about women’s health, you have NO right to regulate abortion or birth control”—never showed up on her profile or in her feed.

Several Truth Social users have also reported that their accounts were permanently suspended after they posted about the hearings of a House committee investigating the assault on the Capitol. In a statement given to the , a spokesperson for the platform said the allegations were “transparently stupid.”

Trump Media & Technology Group didn’t respond to a request to discuss content moderation; it’s disclosed little about its approach aside from saying it works with Hive, a startup that analyzes posts using artificial intelligence. Unlike some larger social networks, Truth Social hasn’t developed a straightforward way for independent researchers to discern what’s happening on the platform overall. Researchers say its content moderation decisions in general have been inconsistent.

By removing “politically dissonant” posts, Lukito says, Truth Social has shown it’s willing to play a heavier hand in content moderation than peers such as Parler and Gab.

Truth Social has said it’s also tried to use content moderation to achieve another aim: cultivating a wholesome platform. Nunes told Fox News in January that Truth Social wants to be “the most family-friendly site.” Using Hive’s technology, it screens for posts that contain sexual images, profanity, threats of violence, and hate speech, says a person with knowledge of the matter, who asked not to be named discussing private business matters. The platform has taken a particularly hard line when it comes to sexualized content. But its practices are resulting in some seemingly innocuous posts getting filtered as “sensitive content,” forcing a user to click past before viewing. Images of the cartoon character Homer Simpson in his underwear, women in bikinis, and clearly nonsexual content have all ended up behind the sensitive content filter in recent weeks.

Young social networks quickly discover how difficult content moderation is, according Katie Harbath, a former public policy director at Facebook. “The devil gets into the details nearly immediately,” she says, “and it seems like that’s what they’re running up against.”

Once a network like Truth Social promises an expansive approach to free speech, any content moderation decisions that run counter to that promise are likely to frustrate people quickly. “Can anyone tell me why this pic is covered with the ‘Sensitive Content’ filter?” wrote one user, referring to an image of a shirtless Jesus on the cross. “Are we back in Fascistbook?”

More stories like this are available on bloomberg.com

©2022 Bloomberg L.P.