WhatsApp is adding a new layer of protection that could stop some scams before users even respond.
WhatsApp is rolling out Scam Alert, an optional feature that uses an on-device machine learning model to identify potential scams in messages from people who are not in a user’s contacts. The feature is currently available in a limited beta as Meta tests it with security researchers and gathers feedback before a broader launch.
The model looks for linguistic and conversational patterns associated with known scams. If it detects a likely scam, WhatsApp displays a warning that only the recipient can see. Users can then block or report the sender, continue the conversation, or mark the chat as trusted.
If a warning is wrong, marking the conversation as trusted removes the alert and prevents Scam Alert from flagging that chat again. Users can also voluntarily share the last five messages from a trusted conversation with WhatsApp to help improve the model.
Privacy remains central
Meta says Scam Alert was designed to work without giving WhatsApp access to users’ encrypted conversations. The detection model runs locally on the device, and message content is not automatically sent back to the company.
The company says it also cannot use the system to send a particular AI model to an individual user selectively. Each model version is recorded on a third-party transparency ledger before distribution, while devices verify published signatures and file hashes before loading a model.
WhatsApp also collects limited performance data, such as how often warnings appear and what users do afterward. According to Meta, those figures are aggregated and protected using confidential computing and differential privacy rather than exposing individual conversations.
Users will eventually be able to inspect Scam Alert activity in WhatsApp under Account > Request Info > Scam Alert Activity, including which messages were analyzed, the result, and the model version involved.
Must-read security coverage
A response to a growing scam problem
The feature arrives as scammers increasingly use messaging platforms to build trust, pressure victims and solicit money or personal information. The Federal Trade Commission said consumers reported $2.1 billion in losses from social media scams in 2025, including $425 million tied specifically to WhatsApp, according to CNET.
That makes the phone-based approach particularly significant: WhatsApp is trying to add an automated layer of protection without requiring its servers to inspect private conversations.
The tradeoff behind the warning
Scam Alert could give users a useful second opinion before they respond to an unfamiliar sender, but it is not a guarantee that a message is safe or fraudulent. Machine learning can miss sophisticated scams or incorrectly flag legitimate conversations.
Scam Alert remains an experiment, not a finished security feature. Meta says it will continue testing the system with security researchers, expand bug-bounty coverage, and refine the model before deciding whether to make it broadly available.
If the approach works, WhatsApp could gain an additional layer of scam protection without abandoning the privacy model that has defined the service — but users will still need to treat any automated warning as guidance rather than proof.
Other News: DentaQuest has disclosed a data breach affecting nearly 15 million people, with exposed information potentially including names, Social Security numbers, and health insurance details.
