More than half of Americans reported receiving at least one scam call per day in 2024. To combat the rise of sophisticated conversational scams that deceive victims over the course of a phone call, we introduced Scam Detection late last year to U.S.-based English-speaking Phone by Google public beta users on Pixel phones.
We use AI models processed on-device to analyze conversations in real-time and warn users of potential scams. If a caller, for example, tries to get you to provide payment via gift cards to complete a delivery, Scam Detection will alert you through audio and haptic notifications and display a warning on your phone that the call may be a scam.
So far as I know, if your device uses their Gemini Nano LLM, it doesn’t reach back to their servers at all unless you OPT IN to the ‘Help service inprove’.
This feature though and a few other calling features has made me switch from iPhone single handedly, I was receiving 6-10 spam calls a day, now I see none because they’re screened in the background. It’s fantastic. I’m hooked on these Pixel features and only hope more move to becoming on device features with the ability to opt in to sending certain things off device.
So, I have several legacy Google Assistant compatible devices that do not work with Google’s new AI. As a result I haven’t switched over to Gemini for pretty much anything and I probably won’t. I’m currently building a Home Assistant system to take the place of Google Assistant when it finally sunsets but the going is slow (I have limited time to dedicate to that specifically at the moment). But for phone specific use, I’m taking the wait and see approach.
Home Assistant is such an awesome tool. I use it every day and shamefully have it linked to my Google Home so Gemini can turn on and off devices when prompted. Aside from that, I could just go the route of setting up a local LLM on my server and having Home Assistant be my new assistant on device so it doesn’t use Google at all.
I definitely recommend Home Assistant though, between the iPhone users and now myself on Android in our home, it makes everything appear native to the end user. Now I just use Zigbee and Zwave devices for everything since they’re more reliable and much cheaper.
My biggest issue right now is trying to figure out what I need/want to work vs things I don’t need. This list is one I’ve been keeping of things that I want it to do/be compatible with:
Weather Calendar News/RSS Feed Light Panel Media Panel Search Query Panel Use of Voice controls Singular touchscreen hub and android phone Works with Google home max speakers and Google nest mini speakers Chromecast equivalent functionality
It’s based on the things I use Google Assistant for daily or at the very least weekly. Most all of it can be done with an android tablet and my raspberry pi. But implementing it to be the way I want isn’t as simple (hasn’t been as simple for me) and I think that’s down to following guides for a lot of things that weren’t necessarily intended to work together cohesively.