For several years now, carriers have been using lists to warn users of potential spam and scam calls. These systems are not foolproof. So what happens when users pick it up? At the Google I/O 2024 developer conference on Tuesday, Google previewed a feature that supposedly warns users about potential scams during calls. .
The feature will be built into future versions of Android and uses Gemini Nano, the smallest version of Google's generative AI service that can run completely on-device. This system effectively identifies “conversation patterns commonly associated with fraud” in real time.
Google cites an example of someone impersonating a “bank representative.” Common scammer tactics such as password requests and gift cards also trigger the system. Although it is well understood that these are all ways to get money out of you, there are still many people in the world who are vulnerable to this type of scam. Once launched, a notification will pop up stating that the user may fall prey to unpleasant characters.
A specific release date for this feature has not been determined. As with many of these things, Google is previewing just how much Gemini Nano will be able to do in the future. However, we know that this feature is opt-in.
That's good. With Gemini Nano, the system will no longer automatically upload to the cloud, but the system will still effectively listen to conversations. This is the sort of thing that stands on the back of privacy advocates' necks.
However, opting in may also mean that some people who could benefit most from such features may not check that box.
Publish an AI newsletter. Sign up here to start receiving it in your inbox on June 5th.
