Deepfake scam ads using AI target seniors on Meta platforms
Scammers have exploited Meta's ad platform to run 63 fraudulent campaigns spending $49 million, using AI-generated deepfakes of U.S. politicians to promote fake government benefits, primarily targeting elderly users. The Tech Transparency Project identified the scams, which Meta removed only after many had run for extended periods.
What happened
Tech Transparency Project identified 63 fraudulent advertisers spending $49 million combined on Facebook and Instagram to run ads using AI-generated deepfakes featuring U.S. political figures promoting false government benefits.
The ads targeted tens of thousands of users, often elderly people, promoting fake stimulus checks, public spending cards, or medical care payments. One deepfake video of Donald Trump falsely promised stimulus checks to Americans over 65 in more than 20 states, with words that did not match Trump's official transcript from April.
Reported impact
Meta deactivated 35 commercial accounts but only after they had published dozens to hundreds of ads each, with six accounts spending over $1 million before removal. Nearly half of the 63 fraudulent advertisers continued promoting the scams until the day before the report.
Index analysis
Tech Transparency Project said scammers were exploiting advances in AI technology, public confusion about social security programs, and lax content moderation on Meta to target new victims.
Meta's political ad authorization process requires advertisers to submit official identification like a driver's license and a U.S. postal address, yet fraudulent advertisers bypassed these controls. Meta removed over 150,000 political ads in the last 12 months for violating its policies, though scam ads continued to proliferate.
Governance controls
Meta stated it would invest in developing new technical defenses as scammers constantly evolve tactics to evade detection. Meta's terms of service prohibit scams and claim the company invests in prevention, though TTP said Meta allowed the activity to persist despite these policies.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.