Deepfake Ads Impersonating Political Figures Promote Fraudulent Benefits on Meta Platforms
A nonprofit watchdog group identified 63 scam advertisers spending $49 million on Meta platforms to run deepfake videos of political figures including Trump, Musk, and prominent Democrats, promoting fictitious government benefits and stimulus checks primarily targeting seniors. Meta approved these ads through its political authorization process and only removed them after they had run extensively and spent substantial sums.
What happened
The Tech Transparency Project identified 63 scam advertisers that collectively spent $49 million on Facebook and Instagram, running more than 150,000 political ads that used deepfake videos of political figures including President Donald Trump, Elon Musk, Alexandria Ocasio-Cortez, Bernie Sanders, and others.
The ads falsely promoted fictitious government benefits including stimulus checks, government spending cards, and Medicare payments. One advertiser, Relief Eligibility Center, ran an ad in April and May featuring a deepfake video of Trump promising stimulus checks, directing users to a website to obtain a 'FREE $5,000 Check from Trump.'
Reported impact
The ads reached tens of thousands of Meta platform users and often targeted seniors and vulnerable groups. The ads gave the impression of offering official Medicare and Social Security benefits, exploiting public confusion around social safety net programs.
Index analysis
Meta's political ad authorization process, which requires advertisers to submit an official ID and U.S. mailing address, failed to prevent the scam ads from being approved and run for days or weeks before removal. Meta approved and allowed these 63 scam advertisers to run their campaigns despite policies prohibiting scams.
Governance controls
Meta removed advertisements from all 63 scam advertisers within the past 12 months for violating policies. Meta disabled 35 ad accounts, though only after they had run dozens or hundreds of ads, with six accounts spending over $1 million before being disabled or deleted.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.