TikTok fails a test of its ban on political ads

Less than three weeks before the presidential election, TikTok is still struggling to consistently enforce its long-standing ban on political advertising, a new report finds.

In September, researchers from the nonprofit Global Witness tested the political ad moderation systems of three leading social media platforms – TikTok, Facebook and YouTube – by submitting ads that contained election disinformation. Examples included ads falsely warning that citizens must pass an English language test to vote; claiming that Donald Trump is ineligible to run for president due to his felony convictions; and calling for a repeat of the Jan. 6, 2021, attack on the U.S. Capitol.

Though the three platforms differ in their rules for political advertising, the ads were designed to violate the policies of all three, the researchers said. Global Witness shared its findings with the Tech Brief before publishing them Thursday.

Facebook and YouTube performed relatively well, though there were caveats in each case. TikTok, the only one of the three that prohibits political advertising altogether, did not.

Global Witness’s researchers submitted the same eight ads to each platform, according to the report. They did not declare the ads as political ads and did not undergo any identity verification processes. To mimic how a bad actor might use “algospeak” to evade detection, they lightly masked some of the ads’ key words by substituting numbers and symbols for certain letters. If the platforms approved any of the ads, the researchers deleted them before they were published so that the misinformation wouldn’t be shown to real users.

Of the eight ads, TikTok rejected only four, despite a policy in place since 2019 that bars political advertising on the platform. It rejected those that mentioned a candidate by name, while approving the ones that didn’t, including ads aimed at voter suppression.

“We were pretty surprised to see that TikTok was doing so badly,” said Ava Lee, Global Witness’s campaign lead for digital threats, adding that the ads it approved were “pretty seriously bad.” The report noted that “in real life, disinformation is rarely as clear as this.”

In a similar test by Global Witness and Cybersecurity for Democracy ahead of the U.S. midterm elections in 2022, researchers found that TikTok approved 90 percent of election disinformation ads. This year, the platform approved misleading political ads for publication in Ireland ahead of European Union elections.

Nor is the problem limited to ads submitted by researchers. In September, NBC News found that 52 politically themed ads running on TikTok in apparent violation of its policies.

As for Facebook, it approved one out of the eight ads in Global Witness’s latest test – one claiming that only people with a valid driver’s license can vote.

“While this report is extremely limited in scope and as a result not reflective of how we enforce our policies at scale, we nonetheless are continually evaluating and improving our enforcement efforts,” Meta spokesperson Corey Chambliss said. “Protecting the 2024 elections online is one of our top priorities, with around 40,000 people globally working on safety and security and more than $20 billion invested in these areas since 2016.”

YouTube did not approve any of the ads. It rejected four of the eight out of hand and notified the researchers that their account was paused until they provided further verification of their identity.

Because the researchers declined to do so, the other four ads were never reviewed. Global Witness’s Lee noted that it was unclear whether some of them might have been approved had they provided identification.

“Our enforcement is multilayered and we rely on a variety of tools to combat abuse on our platform, including policies against hate speech and misinformation, advertiser verification requirements, and signals that help us detect suspicious accounts,” Google spokesperson Michael Aciman said. “These measures helped us appropriately block this account from running ads.”

The results of Global Witness’s simple test are useful, but they shouldn’t be taken as definitive verdicts on the companies’ respective systems.

“The report is helpful in indicating how the platforms are doing in determining if an ad is political and needs to go through verification and disclaimer process,” said Katie Harbath, CEO of the tech consultancy Anchor Change and a former Facebook public policy director. That’s true, she added, “especially for a platform like TikTok that doesn’t allow political ads and is a place where a large number of people are getting their news and information.”

“What it doesn’t show,” she said, “is how good the platforms are in then determining if the content of the ad violates their terms, because they never ran and many platforms don’t prescreen ads for content violations before they run.”

“Overall, this shows the ongoing challenge and gap of what tech companies’ policies are and how well they can enforce them,” Harbath said.