Android users will soon get an AI scam call detection feature, but there’s a catch

Call-based scams are on the rise and we’ve seen their technology alarmingly improve over time. From fraudulent salesmen and innocent-sounding telecallers to imposters and fake bank executives, no one person can mark themselves safe from a scammer today. In fact, we could even go as far as claiming that three out of five calls every day are spam and are often always plaguing us on a busy afternoon. 

While various third-party apps today are equipped to block incoming calls and messages, they usually rely on user data and flag a caller suspicious only after a large number of people have reported it. But all that is likely to change with an AI spam detection feature which was showcased last night at Google I/O 2024.

This upcoming feature will give you real-time updates during a call and alert you almost instantly if it suspects a fraudulent caller. But, of course, as it is with all good things, there’s a catch.

Google AI spam detection for Android: How exactly will this feature work?

According to Google, the upcoming feature will utilise Gemini Nano, a reduced version of Gemini LLM for Android devices. It will typically identify any patterns associated with scams and alert a user during an ongoing call. Some examples include scammers posing as bank executives and asking for sensitive information, such as OTPs or PIN numbers. 

ALSO READ: OpenAI announces a major ChatGPT privacy update; here’s what you should know

The pop-up displayed at Google I/O had text that read ‘Possible bank scam: Banks will never ask you to move your money’. The feature will have a notification sound to let the user know when the message pops up. However, apart from this, it will also give users the option to end the call. 

So, what's the catch?

For Google to give you real-time updates about your call, you will also have to extend permissions to the company and allow it to listen to your conversations. Gemini AI will silently run in the background to analyse spoken words or phrases during a call, but this will also mean it cannot be switched off for regular calls. 

While this is certainly a useful feature to have at a time when call scams are on a rise, it would also mean that at some level you may not be able to have entirely private conversations.

ALSO READ: Google will soon put a stop to bulk email spamming; here’s what you should know

A Google representative, in conversation with CNET, however mentioned how user data will stay private on their smartphone. Google is yet to announce a potential launch date or a roll-out schedule, and as of today, it also remains unclear as to which smartphones will make the cut. 

The concept, for now, has sparked various discussions and there doesn’t seem to be a way to exercise more granular control during calls. However, with the right regulations in place, it may just be a step in the right direction that harnesses artificial intelligence for its greater good.