Google Chrome to use on-device AI to detect tech support scams

Google is introducing an AI-powered anti-scam capability in Chrome 137 that uses the on-device Gemini Nano model within Enhanced Protection to analyze pages locally for signals of tech support scams—such as fake virus warnings and full-screen lockouts—and, when flagged, sends metadata to Safe Browsing for verification before warning users. The approach emphasizes privacy and performance through local processing, token limits, asynchronous operation, and GPU throttling, will be enabled by default for users who opt into Enhanced Protection, and is slated to expand to detect additional scam types and reach Android in 2025.