Meta acts on 3.7 million accounts with Singapore Police to fight scam networks

Meta announced it took enforcement action on 3.7 million accounts, pages, and pieces of content in partnership with the Singapore Police Force, in a coordinated crackdown on organized scam networks. The targeted operations ran across borders and platforms, pushing fraudulent stock tips, fake brands, and guaranteed-return investment schemes — a category of fraud that has surged alongside generative AI's ability to mass-produce convincing scam content.

The scale (3.7 million actions) reflects both the size of the problem and Meta's reliance on automated detection: identifying coordinated inauthentic behavior at that volume requires AI-driven classifiers to spot patterns across accounts, content, and payment signals, then human and law-enforcement partnership to act.

The partnership model with a national police force is the notable structural element — platform enforcement paired with state investigative power against transnational criminal networks that no single actor can dismantle alone. It positions Meta's trust-and-safety operation as a public-private enforcement partner, useful reputationally amid ongoing scrutiny of its platforms.

The context is a broader AI double-edge: the same generative tools Meta ships in Muse and Llama also lower the cost of producing scams, making AI-powered detection a necessary counterweight. The caveat, familiar from prior takedown announcements, is that raw action counts don't reveal precision (false positives) or durability (whether networks simply respawn). What to watch: whether the Singapore partnership becomes a template Meta replicates with other national authorities, and whether AI-generated scam content — deepfakes, synthetic endorsements — features prominently in future enforcement reports.