Android enhances scam defense in Google Messages using cloud-based Gemini AI

Android is rolling out an enhanced protection setting in Google Messages that utilizes cloud-based Gemini AI models to evaluate the intentions of potential scammers in conversations with unrecognized senders. This feature augments existing on-device detection, allowing identification of novel attacks, while user data remains encrypted and processed in a secure dedicated space. The enhanced protection mode is opt-in and also uses on-device Gemini Nano AI to inspect real-time web signals in unfamiliar URLs before they are opened. The scam detection feature, previously on Pixel phones, is expanding to devices from Honor, Motorola, Oppo, OnePlus, Realme, Tecno, Vivo, Xiaomi, and more, with Google planning to reach hundreds of millions of additional users later this year. Additional security features include blocking apps from enabling call forwarding and automatically concealing sensitive information during screen-sharing in Chrome and Android WebView.