Published • loading... • Updated
WhatsApp’s New Scam Alert Feature Will Flag Scam Messages While Keeping Your Chats Private
The optional beta feature scans messages from noncontacts on-device and lets users block or report suspected scams without sending chat content to Meta.
- Meta announced Scam Alert on Wednesday, an optional on-device tool for WhatsApp designed to flag suspicious messages from non-contacts. The feature aims to protect the app's over 3 billion daily users from fraud.
- The Federal Trade Commission reported that consumers lost $2.1 billion to social media scams in 2025, with Americans losing $425 million on WhatsApp alone. These figures reflect persistent threats scammers pose to the platform.
- Once enabled, Scam Alert downloads a machine learning model to the device that "runs inferences to classify whether incoming messages from non-contacts match known scam patterns." This analysis maintains end-to-end encryption, as no content leaves the user's phone.
- If triggered, the warning displays only to the recipient, who can block, report, or continue chatting. "If they decide that a warning is incorrectly flagged," users can mark the chat as trusted to remove future alerts.
- Currently in limited Beta, Scam Alert represents a new layer in Meta's "Swiss Cheese" security model. The company is working with the WhatsApp bug bounty community to stress-test the feature before broader rollout.
Insights by Ground AI
12 Articles
12 Articles
WhatsApp's On-Device AI Scam Alert Tests Privacy Limits Against Sophisticated Voice Clones
Scammers stole $1.27 million from a man in Hong Kong. They used AI-generated voice notes on WhatsApp that sounded exactly like his father. The messages carried urgency. They demanded immediate transfers. He sent the money. Repeatedly. That incident, reported this month, underscores a harsh reality. Familiar voices no longer guarantee safety. Criminals now clone them with just seconds of audio. Yet Meta has a response in limited testing. An on-de…
Meta has given WhatsApp tool so that it can protect users. The most recent has arrived for testing and ensures that messages that try to mock or deceive who uses this service...
·Lisboa, Portugal
Read Full Article
Coverage Details
Total News Sources12
Leaning Left1Leaning Right1Center5Last UpdatedBias Distribution72% Center
Bias Distribution
- 72% of the sources are Center
72% Center
14%
C 72%
14%
Factuality
To view factuality data please Upgrade to Premium
