Google I/O 2024 brought forth a huge array of new ideas and features from Google, including some useful life hacks for the lazy, and AI was present almost everywhere you looked. One of the more intriguing new features can detect scam calls as they happen, and warn you not to hand over your bank details – and it’s not as creepy as it might at first sound.

The feature works like this: if you get a call from a scammer pretending to be a representative of your bank, Google Gemini uses its AI smarts to work out that the impersonator is not who they claim to be. The AI then sends you an instant alert warning you not to hand over any bank details or move any money, and suggests that you hang up.

Involving AI in the process raises a few pertinent questions, such as whether your bank details ever get sent to Google’s servers for processing as they're detected. That’s something you never want to happen, because you don’t know who can access your bank information or what they might do with it.

Fortunately, Google has confirmed (via MSPoweruser) that the whole process takes place on your device. That means there’s no way for Google (or anyone else) to lift your banking information off a server and use it for their own ends. Instead, it stays safely sequestered away from prying eyes.

Google’s approach cuts to the core of many concerns about the growing role AI is playing in our lives. Because most AI services need huge server banks to process and understand all the data that comes their way, that means you can end up sending a lot of sensitive data to an opaque destination, with no way of really knowing what happens to it.

This has had real-world consequences. In March 2024, AI service Cutout.Pro was hacked and lost the personal data of 20 million users, and it’s far from the only example. Many firms are worried that employees may inadvertently upload private company data when using AI tools, which then gets fed into the AI as training data – potentially allowing it to fall into the hands of users outside the business. In fact, exactly that has already happened to Samsung and numerous other companies.

This all goes to show the importance of keeping private data away from AI servers as much as possible – and, given the potential for AI to become a data privacy minefield, Google’s decision to keep your bank details on-device is a good one.

• The growing threat of data breaches in the age of AI and data privacy
• The 7 biggest AI announcements from Google I/O 2024 – from Gemini to Project Astra
• CISOs are nervous Gen AI use could lead to more security breaches