Documents seen by Reuters indicate Meta projected about 10% of its 2024 revenue — roughly $16 billion — could come from ads tied to scams or prohibited goods. Another internal note estimated users are exposed to 15 billion high-risk scam ads each day across Facebook, Instagram and WhatsApp, a figure that dwarfs many countries’ entire ad markets. The cache spans finance, safety and lobbying teams, suggesting leadership weighed enforcement costs against business impact.

Executives crafted a system that bans an advertiser only when automated models hit a 95% fraud certainty, the papers say. When suspicion falls below that threshold, Meta applies “penalty bids,” charging suspected scammers more for inventory to discourage them — and, yes, to recoup some value. Users who click these ads are then more likely to see similar ones, a personalization side-effect the files openly acknowledge.

Critics say the approach normalizes risk and shifts the burden to the public. A fraud examiner told Reuters regulators wouldn’t tolerate banks profiting from fraud and shouldn’t tolerate it in tech either. Meta counters that the 10% figure was rough and overstated, and that scam reports fell 58% globally over 18 months, with 134 million scam ad items removed in 2025.

Read more: full story