Google has unveiled AI-powered security features for Android devices, introducing real-time scam call detection and enhanced malware monitoring. The new capabilities leverage on-device AI processing to protect users from fraud and malicious applications without routing sensitive data to external servers.
AI-Powered Scam Call Detection
The scam detection system is powered by Google’s Gemini Nano AI model, running entirely on-device. It analyzes voice calls in real time, identifying patterns associated with common fraud tactics: impersonation of banks or government agencies, artificial urgency, and requests to transfer money or share security codes. All call audio is processed locally — no call data leaves the device. The feature must be manually activated through the Phone app settings, giving users explicit control over whether it operates.
Enhanced Google Play Protect with Live Threat Detection
The update introduces Live Threat Detection, a significant upgrade to Google Play Protect. Rather than scanning only at install time, this feature continuously monitors running application behavior — analyzing how apps use system permissions, interact with other apps, and access sensitive data. Processing occurs locally through the Android Private Compute Core, an isolated environment that prevents the monitoring data from being accessed by other apps or sent to Google’s servers.
Initial Focus: Stalkerware
The first threat category targeted by Live Threat Detection is stalkerware — malicious applications designed to silently track a device’s location, messages, and calls without the owner’s knowledge. Future updates will expand detection to adware, spyware, and banking trojans.
Who Is Affected and Device Eligibility
The initial rollout targets Google Pixel 6 and newer devices. These phones include the Tensor chip required to run Gemini Nano on-device at sufficient speed for real-time call analysis. Other Android manufacturers are expected to gain access in later phases as Google expands hardware requirements. The scam call detection feature currently supports English only; multilingual support is planned for future updates.
Users most at risk from the threats these features address include:
- Older adults who are frequently targeted by phone-based financial fraud
- Android users who have previously installed apps from outside the Play Store
- People in relationships involving intimate partner surveillance, where stalkerware is commonly deployed
Enabling Google AI Security Features on Android Devices
To take advantage of these protections, Android users should:
- Enable scam call detection manually in the Phone app under Settings > Scam Protection (Pixel 6+ running Android 12 or later required)
- Verify that Google Play Protect is active: open the Play Store, tap your profile icon, select Play Protect, and confirm scanning is enabled
- Keep Android OS and Google Play Services updated — Live Threat Detection requires up-to-date system components
- Review installed app permissions in Settings > Privacy > Permission Manager and revoke any that seem excessive for the app’s stated function
- Consult the Android Security documentation for details on Private Compute Core and its privacy guarantees
