TikTok and Facebook approved advertisements containing blatant US election falsehoods just weeks ahead of the vote, a watchdog investigation revealed Thursday, calling into question the tech platforms’ policies to detect harmful disinformation.

The advocacy group Global Witness submitted eight ads containing false election claims to the Chinese-owned video-sharing app TikTok, the Meta-owned Facebook, and Google-owned YouTube to test their ad systems in the final stretch of the 5 November election.

The ads carried outright election falsehoods — such as people can vote online — as well as content promoting voter suppression, inciting violence against a candidate, and threatening electoral workers and processes.

TikTok “performed the worst,” Global Witness said, approving four of them despite its policy that prohibits all political ads.

Facebook approved one of the ads submitted.

“Days away from a tightly fought US presidential race, it is shocking that social media companies are still approving thoroughly debunked and blatant disinformation on their platforms,” said Ava Lee, the digital threats campaign leader at Global Witness.

The study comes as researchers warn of the growing perils of disinformation — both from domestic actors and foreign influence operations — during a tight election race between the Democratic contender, Vice President Kamala Harris, and Republican nominee Donald Trump.