Meta's Failing Ad Moderation: Health Scams Targeting EU Users

AI Forensics analyzed roughly 450 million advertisements drawn from Meta's Ad Library API since August 2023, using pattern-matching to detect near-identical text republished across many different advertiser pages, then applying semantic analysis to isolate clusters of health-related content likely to be fraudulent.

The analysis identified more than 46,000 ads promoting unapproved drugs or making deceptive health claims, which collectively reached EU users over 292 million times. The investigation also found structural weaknesses enabling the scams: Meta's advertiser "beneficiary" and "payer" disclosure fields are self-reported and unverified, advertisers can display a URL that differs from the actual landing page, and some advertiser accounts impersonated celebrities or medical professionals without detection. Many flagged campaigns violated more than a dozen of Meta's own advertising and community standards, and activity was found to be ongoing into 2025.

The report concludes that these gaps reflect a failure to implement adequate risk-mitigation measures required under the Digital Services Act's Article 34, rather than isolated moderation errors.