Meta is rolling out an optional feature called Scam Alert on WhatsApp that uses on-device artificial intelligence to flag suspicious scam messages, The Verge reported. The new feature is arriving in a limited beta test phase.
Earlier this year, Meta also launched scam detection for device-linking requests on WhatsApp. The new feature is another layer of protection against fraud in the app.
How it works
When the model identifies a conversation as a likely fraud attempt, the user sees a warning inside the chat, which is not visible to the other person. From there, it's possible to block the contact, report the conversation, or continue the dialogue.
If the user believes the warning was incorrect, they can mark the chat as trusted. In that case, the alert is removed and the tool will not flag that conversation again. Those who mark a chat as trusted can also choose to share the last five messages received with WhatsApp to help improve the feature's accuracy.
WhatsApp has more than 3 billion users and is one of the areas with the highest scam activity, such as bank transfers and frauds known as "pig butchering." The FTC (U.S. Federal Trade Commission) reported that victims reported losses of US$ 425 million in scams on WhatsApp alone, out of a total of more than US$ 2.1 billion lost to social media scams in 2025.
Meta said that no message content leaves the user's device for classification when the feature is enabled, and that nothing is automatically reported to WhatsApp, Meta, or third parties. Users can disable Scam Alert at any time.
