WhatsApp is trialling a new on-device AI system that detects suspicious chats from unknown contacts, aiming to bolster scam prevention without compromising end-to-end encryption amid rising messaging fraud.
WhatsApp is trialling a new scam warning system that uses on-device artificial intelligence to flag suspicious chats from unknown contacts, in a move designed to improve safety without weakening end-to-end encryption. According to the company, the feature analyses conversations on the phone itself rather than sending message content to Meta’s servers, and it is being rolled out in a limited test to selected users.
Meta said: “No message content leaves the device for classification or is auto-reported to WhatsApp, Meta, or anyone else.” The system downloads a compact AI model to the handset after activation and watches for patterns associated with common frauds, including fake offers and payment requests. If it detects a possible scam, a warning appears within the chat and only the recipient can see it.
The alert is optional during testing, and WhatsApp is giving users several choices when one appears, including blocking the sender, reporting the conversation, continuing to chat or marking the contact as trusted. Gadgets 360 reported that a separate trust-warning screen is also being tested for chats from unfamiliar numbers, with the sender unaware of the user’s decision. WhatsApp’s approach follows a broader push by Meta to strengthen anti-scam defences across its apps, including earlier safety tools introduced for WhatsApp last year and additional detection measures across Facebook and Messenger in March.
The limited rollout also reflects the difficulty of automated scam detection: while the system is designed to catch fraud early, it may also flag legitimate first-time messages from businesses or people the user does not know. That matters in markets such as India, where WhatsApp is widely used for payments and where scammers often rely on urgency and impersonation. Recent cases, including a Hong Kong fraud in which criminals reportedly used AI voice notes to impersonate a family member, underline how quickly messaging scams are becoming more convincing.
Disclaimer: This article is intended to inform and educate, not to recommend or endorse any financial product, investment or strategy. Please consider your own financial circumstances and seek professional advice where appropriate before making financial decisions.
