WhatsApp has enabled a new "anti-scam" feature: how to avoid falling into artificial intelligence traps

WhatsApp , the most popular mobile messaging service, has taken note and addressed a recurring request from Argentine users: an end to digital scams . To this end, it announced new features to detect and stop scams circulating on its platform, often with messages promising quick cash or rewards for simple tasks.

These measures come at a key moment for the technology company Meta, in a context where cybercriminals' criminal campaigns are increasingly sophisticated and indecipherable , using artificial intelligence (AI), leaked data, and emotional manipulation and social engineering techniques to deceive thousands of users.

According to cybersecurity firm Kaspersky , the fraudulent schemes detected on WhatsApp are becoming increasingly sophisticated, something that WhatsApp has recently noted and has taken tough action against, many of them in response to user complaints.

The platform has removed more than 6.8 million fraudulent accounts so far this year, most of which were linked to internationally organized scam centers.

Initial contact often occurs on social media, dating apps, or via text message, and the user is then directed to WhatsApp, where they receive AI-powered messages promising remote work, cryptocurrency investment opportunities, fake parcel deliveries, social aid campaigns, or payments for simple tasks like liking TikTok.

“The automation of attacks and the trust users have in WhatsApp make it easy for scams to spread ,” explains Fabio Assolini, director of Kaspersky’s Global Research Team for Latin America.

Most of these messages are designed with artificial intelligence to sound credible. In fact, WhatsApp pointed to an example where it collaborated with OpenAI , the company that owns ChatGPT, to dismantle a cyberscam center in Cambodia: scammers used ChatGPT to generate an initial message in several languages with a link to a chat and a fake offer.

Online scammers even use real user data to achieve their goal: names, ID numbers, or online shopping history , making them even harder to detect.

Meta introduced a new safety feature called “Group Safety View.” This feature alerts users when an unknown contact adds them to a group , displaying the message: “You can leave the group without even checking the conversation.”

This means you no longer need to open the chat to leave. Notifications will also be muted until the user decides whether to stay in the group or leave.

The app is also testing "Individual Chat Alerts." These are mechanisms that alert users when they start a conversation with someone who isn't in their contacts. These alerts will provide more context about who the contact is , helping them decide whether to continue the conversation.

Despite these features, the user's role and the control they exercise over their contacts and conversations must be strict. To avoid becoming a victim, WhatsApp advises that you should never respond to messages immediately . Furthermore, hesitating when faced with a suspicious request—a third party pressuring you to send money or provide a PIN—is key.

It's also important to verify who is initiating the contact , at least through another contact method outside the app, since the scammer often changes their identity to someone close to the victim.

For cybersecurity experts, such as Kaspersky, these actions are a step in the right direction; although they warn that in Latin America, their effectiveness may be limited , as many scams are personalized and do not involve large volumes of mass messages, making automatic detection difficult.