Meta Launches AI-Powered Scam Alert to Detect WhatsApp Fraud Attempts

Meta has introduced a new feature called Scam Alert on WhatsApp to combat the growing problem of scam messages. The tool uses artificial intelligence to identify suspicious patterns in conversations that may indicate fraud attempts and warns users in real time during chats. When a suspicious message is detected, an alert appears within the chat, allowing users to block the sender, report them, or continue the conversation. Users can also mark the chat as safe if the alert is a false positive, which removes the warning.

Importantly, Meta states that the AI model operates directly on the user's device, ensuring that message content is not sent to Meta’s servers. The system analyzes conversation structure and text patterns based on previously reported scam cases. Data collected to evaluate the feature’s effectiveness is anonymized and statistical, preventing identification of individual users. The Scam Alert feature is not yet available to all users and can be disabled if desired.

This development comes amid significant financial losses due to WhatsApp scams, with the U.S. Federal Trade Commission reporting $425 million stolen via WhatsApp scams in the United States alone. Meta’s new tool aims to provide timely warnings that could prevent users from falling victim to costly fraud.