Meta Rolls Out On-Device AI Scam Detection for WhatsApp
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Meta is rolling out an optional Scam Alert feature on WhatsApp in limited beta, using on-device machine learning to flag suspicious messages and show users a warning inside the chat that remains invisible to the other party.
- Scam Alert is fully user-controlled: people can block, report, continue the conversation, or mark the chat as trusted — and if trusted, they can optionally share the last 5 messages with WhatsApp to improve the feature's accuracy.
- Meta states that no message content leaves the user's device for classification, nothing is auto-reported to WhatsApp, Meta, or anyone else, and users can disable Scam Alert at any time.
- WhatsApp has more than 3 billion users, making it one of the largest areas of scam activity including wire transfers and pig butchering schemes, according to the video.
- The FTC reported victims lost $425 million to scams via WhatsApp alone, out of more than $2.1 billion lost to social media scams in 2025.
- Earlier this year, Meta also launched scam detection for device linking requests on WhatsApp, making the new Scam Alert the company's second such initiative on the platform in 2025.
Why it matters: With the FTC reporting $425 million lost to WhatsApp scams alone in 2025 and WhatsApp's 3 billion users making it a massive fraud target, Meta's on-device approach tries to scale detection without compromising the end-to-end encryption stance that WhatsApp is marketed on — processing happens locally and nothing is auto-reported, so users gain a warning layer without giving Meta visibility into their chats.
Ask SkimNews
