WhatsApp is working on a new security feature that could help users identify potentially fraudulent conversations before they fall victim to a scam.

The messaging platform is reportedly testing an AI-powered Scam Alert feature that can examine conversations and look for signs of suspicious activity. If a chat appears to follow patterns associated with scams, WhatsApp could display a warning directly within the conversation.

The feature is currently being tested with a limited group of beta users and is expected to remain optional.

How will WhatsApp's Scam Alert work?

The proposed feature is designed to use on-device machine learning to assess whether a conversation may be linked to a potential scam.

If suspicious patterns are detected, the user could receive a warning inside the chat. The alert would be visible to the user who receives it, meaning the other person in the conversation would not be notified that a warning has appeared.

After receiving an alert, users could decide what to do next. Options may include blocking the contact, reporting the conversation, or continuing the chat if they believe it is legitimate.

What if WhatsApp makes a mistake?

AI-based detection systems can sometimes flag legitimate activity by mistake. WhatsApp's planned feature reportedly takes this possibility into account.

If a genuine conversation is incorrectly identified as suspicious, the user could mark the chat as trusted. This would help remove the warning and could prevent the system from flagging the same conversation again.

Users may also have the option to voluntarily share their five most recent received messages with WhatsApp. This information could potentially be used to improve the accuracy of scam detection.

Importantly, sharing these messages is reportedly optional.

WhatsApp is focusing on privacy

Privacy is a major part of the proposed system.

According to reports, scam detection would take place directly on the user's device, meaning the content of messages would not need to be sent elsewhere simply to determine whether a conversation appears suspicious.

The appearance of a scam warning would also not automatically mean that the conversation has been reported to WhatsApp or Meta.

Users would reportedly remain in control of the feature and could disable scam alerts if they do not want to use them.

Why could this feature be useful?

Online scams increasingly rely on convincing conversations rather than obvious spam messages. Fraudsters may pose as bank representatives, delivery agents, recruiters, customer-support staff, or even acquaintances.

A warning appearing while a suspicious conversation is taking place could give users an additional opportunity to stop and reconsider before sharing sensitive information, clicking a link, sending money, or revealing an OTP.

However, an AI warning should not be treated as a guarantee that a conversation is safe. Scammers can constantly change their tactics, and legitimate conversations could potentially be flagged incorrectly.

When will the feature be available to everyone?

The Scam Alert feature is reportedly still in the testing stage and has been made available to a limited number of beta users.

WhatsApp has not yet confirmed a full public rollout date. The company may continue refining the detection system based on testing before expanding the feature to more users.

If released widely, the feature could become another layer of protection against scams while keeping message analysis primarily on the user's device.