On May 14, at the I / O 2024 developer conference, Google announced the introduction of AI scam call detection for Android, alerting users to possible scams during calls and encouraging them to end such calls.
The feature is based on the locally running Gemini Nano model, which matches to find fraudulent language and other conversation patterns commonly associated with scams during a call.
For example, if a scammer poses as a bank employee and asks for personal information such as your password or card PIN, or asks for a payment via a gift card, or asks the user to transfer money to them urgently, Gemini Nano recognizes all of the above scenarios and thus alerts the user to the risk.
Security-wise, Google says that these new protections are implemented entirely on the device, so conversations monitored by the Gemini Nano will remain private.
It’s not yet known when this fraud detection feature will be rolled out, but Google says it’s in a disabled state by default and will need to be manually enabled by the user for it to take effect, and that Google will share more information later this year.
