AI-powered deepfakes fuel Pix tax scams on Meta's platforms
A study by Brazil's Federal University of Rio de Janeiro found that Meta's platforms hosted thousands of fraudulent advertisements promoting false Pix tax schemes and government programs, with scammers using AI-generated images of politicians. The research indicates Meta reportedly profited from these ads while fraudulent activity grew 35% after the government revoked the disputed Pix regulation.
What happened
NetLab, a research lab at Brazil's Universidade Federal do Rio de Janeiro, published a study on February 5, 2025, analyzing fraudulent advertisements on Meta platforms. During January 10-21, 2025, researchers monitored 1,770 advertisements with malicious content from 151 advertisers, identifying 87 fraudulent websites used to target users.
Scammers exploited confusion caused by a revoked Receita Federal regulation that would have required financial institutions to report Pix transactions above R$ 5,000 per month for individuals. The fraudulent ads impersonated government institutions and promoted both real and fictitious government assistance programs, using AI-generated images of politicians to lend credibility.
Reported impact
According to the study, fraudulent advertisement volume on Meta platforms grew 35% after the government revoked the Pix reporting rule on January 15, 2025. Scammers exploited the desinformation surrounding financial inclusion policies to defraud Brazilian citizens through Meta's advertising systems.
Index analysis
The research suggests Meta's platforms facilitated fraudulent advertising through inadequate moderation of paid advertisements. In 40.5% of cases, advertisements were placed by actors impersonating the federal government, exploiting public uncertainty about the policy change to promote scam schemes.
Governance controls
Meta stated in response to the study that activities designed to deceive, defraud or exploit others are not permitted on its platforms and that the company is 'always improving technology to combat suspicious activities.' The company recommended users report content they believe violates community standards or advertising policies.
Classification
Sources and evidence
Sources not yet published
This record was researched and written by the Index from original reporting. Its sources are still being verified by an editor, and they are listed here once that check is complete. Until then the record carries no reviewer.