Google is finally rolling out the long-awaited AI-based scam detection feature to its Pixel devices with the latest beta update for the Phone app. First introduced at I/O in May, the feature uses on-device AI models to identify and notify users of potential scam calls.

The scam detection feature can recognize typical conversation patterns in scam calls, such as a caller pretending to be from your bank asking you to transfer funds due to an alleged account breach. It then delivers an audio and haptic alert as well as a visual warning to prevent users from being scammed.

To ensure privacy, Google says the feature handles the call entirely on the device and does not send audio or transcripts of the conversation to the company’s servers. Additionally, the feature is disabled by default and users have the option to enable it in the Phone app settings for a particular call or for all calls.

During its initial announcement, Google claimed that the scam detection feature would be powered by the Gemini Nano On-device AI model. Although this is true for the Pixel 9 Seriesthe company has now confirmed that the feature will rely on other machine learning models on older Pixel devices. It is not immediately clear whether this will negatively affect the speed and accuracy of scam call detection.

Scam detection is rolling out to Pixel 6 and newer models as part of the latest Phone app beta. The feature currently supports English calling and is only available on Pixel devices in the United States. At this time we have no information regarding a wider release or additional language support.