550 TikTok Accounts Sold Fake AI Protests for Cash. Here's the Audit Trail That Could Have Caught Them.

On January 26, 2026, the Spanish nonprofit Maldita.es published an investigation that should alarm anyone working in AI safety, content moderation, or platform governance. Over two months, researchers documented 550 TikTok accounts across 18 countries producing AI-generated videos of protests that never happened — not for political reasons, but to qualify for TikTok's creator monetization program . The scheme was straightforward. Use a VPN to access Sora from Canada. Generate realistic protest footage. Post it to TikTok targeting countries where the monetization program operates (UK, US, France, Germany, etc.). Hit the 10,000-follower threshold. Start collecting revenue — or sell the account. Some creators didn't even bother removing the Sora or Gemini watermarks from their AI-generated videos. The investigation, covered in depth by TechPolicy.Press on February 22, revealed a structural problem that goes beyond content moderation. TikTok's own community guidelines prohibit inauthentic