Published on: November 15, 2024
Google is introducing new AI-powered features for Pixel phones to boost security, with the promise of adding them to Android phones in the near future, too.
The real-time protection features are being rolled out to Pixel 6 or newer devices. The more impressive one is called Scam Detection, and it leverages on-device AI to analyze conversation patterns typically associated with scammers.
Here’s how it works: When you make or receive a call, your Pixel device’s AI actively listens in on the conversation. On the Pixel 9 series, the advanced Gemini Nano model handles this detection, while the Pixel 6 and newer devices use an earlier version of the AI.
The system looks for red flags, such as callers asking you to urgently transfer funds for suspicious reasons, requesting personal account details, or pretending to be from your bank.
If the AI identifies a potential scam, you’ll hear an aduio alarm and vibration, alongside a warning that says, “Likely Scam” and “Suspicious activity detected for this call.” You’ll then have the option to either mark the call as “Not a scam” or choose to “End call.”
The feature operates entirely on the device, meaning Google doesn’t store any conversation locally or on its servers. It’s turned off by default, but you can easily enable it in settings.
Currently, the feature can identify scam calls in English and is being gradually rolled out to public beta users in the United States — though, only those enrolled in the “Phone by Google” beta program and have the latest version of the Phone app installed.
The second useful addition that Google announced is live threat detection within Google Play Protect. Initially, the feature will focus on detecting stalkerware—apps that gather sensitive data without consent. It’ll later expand to identify other types of malicious activities.
This feature looks for apps “that try extra hard to hide their behavior or lie dormant for a time before engaging in suspicious activity,” according to Google.
Like Scan Detection, Live threat detection happens directly on your device, meaning it doesn’t rely on any external servers. The feature uses Private Compute Core, which is a secure Android system designed to protect your data.