Google enhances Android security with Gemini AI-powered scam detection

Google has announced a suite of AI-driven security updates for the Android ecosystem designed to combat digital fraud, including phishing, spoofed calls, and data theft. Utilizing Gemini models, the new protections will be integrated into Google Messages to analyze conversation contexts and identify scammer intent. This includes an opt-in ‘Enhanced Protection’ mode that uses Private AI Compute to process data securely on-device or in encrypted environments.

To protect against malicious links, the system will employ the on-device Gemini Nano model to perform real-time safety checks on unfamiliar URLs, providing users with redesigned, context-rich warnings before they click. Additionally, Android is expanding screen-sharing protections to include Chrome and Android WebView, which will automatically mask sensitive information such as passwords, login fields, and banking details during live sessions.

To address telephony fraud, Google is introducing features to prevent apps from secretly activating call forwarding in the background and to alert users when suspicious call-forwarding codes are entered. In Brazil, the company’s caller identity verification technology has already successfully intercepted over 3 million fraudulent bank calls since June 2026. These features are expected to roll out to various devices, including the Google Pixel 10 series and Samsung Galaxy S26, with broader availability for other manufacturers planned through 2026.