WhatsApp To Roll Out AI Scam Detection: How It Could Protect Users From Fraud

WhatsApp is testing a new optional feature called Scam Alert that uses artificial intelligence to warn users about potential scam messages while keeping their messages protected by end-to-end encryption.

In a post on Meta’s Engineering blog, Meta said scam tactics are evolving from impersonation and social engineering to AI-generated lures, adding that it is working to keep its protections ahead of scammers while protecting users’ personal messages with end-to-end encryption.

How WhatsApp’s AI Scam Detection Feature Works

Scam Alert uses an on-device machine learning model to analyse incoming messages from people who are not saved as contacts. The model is trained on patterns observed in scam conversations from reports that users have sent to WhatsApp. It uses conversational structure and linguistic signals to identify whether messages match known scam patterns.

The feature is optional. Once a user turns it on, the machine learning model is downloaded to the device, where it analyses incoming messages.

Meta said, “No message content leaves the device for classification or is auto-reported to WhatsApp, Meta, or anyone else.”

If the model identifies a message as a likely scam attempt, the user will see a warning inside the chat. The warning will not be visible to the other person.

Users can then choose what to do. They can block or report the sender, or continue the conversation if they believe the warning is incorrect.

If a chat is wrongly flagged, users can mark it as trusted. The warning will then be removed, and Scam Alert will not flag that chat again.

Users Will Remain In Control

Meta said Scam Alert has been designed around three principles: on-device processing, no automatic reporting and user control.

The company said, “Scam Alert is a user-controlled tool that provides additional information to the user, and the user can turn it off or on at any time.”

If a user marks a wrongly flagged chat as trusted, they can also choose to share the last five messages received with WhatsApp to help improve the feature’s accuracy.

WhatsApp will also provide transparency logs that allow users to see which messages were analysed, whether the model flagged them and which model version was used.

When Will WhatsApp’s Scam Alert Be Available

Scam Alert is currently being rolled out on a limited basis in Beta. Meta said it will continue testing and improving the feature before making it available more broadly.

The company has not announced a timeline for making the feature available to all WhatsApp users.

The feature is designed to add another layer of protection against online scams while keeping message content on the user’s device during the classification process.