WhatsApp is rolling out Scam Alert, a new feature powered by on-device machine learning that analyzes incoming messages from unknown numbers for signs of fraud. The system examines linguistic patterns and conversation structure against known scam indicators, flagging suspicious messages before they can deceive you.
When Scam Alert detects a potential fraud attempt, you get options to block the sender, report the message, or mark the conversation as trusted if it turns out to be legitimate. This approach keeps the detection work local on your phone rather than sending data to Meta's servers, protecting your privacy while catching common fraud tactics in real time.
The feature is launching in limited beta on WhatsApp's beta version first, with broader rollout expected to follow. The shift toward on-device AI for safety shows how companies are using machine learning not just for convenience, but as a practical defense against the growing wave of social engineering attacks that plague messaging apps.