Google Is Bringing Gemini to Android Auto, and It Sounds So Annoying

Gemini is coming for every part of your Google-led life, whether you like it or not. It’s even en route to the car, which I definitely do not like. Once it officially rolls out, I hope we’ll understand how to use Gemini from behind the wheel. The current demonstration of what it’s capable of doesn’t instill much hope in me.

The sleuths over at Android Authority dove into some code and activated the dormant Gemini capability tucked inside a build of Android Auto. In the video, which you can watch on YouTube, the site asks Gemini to plan a trip within a budget. Its answer is super generic and distinctly unhelpful for someone steering a vehicle. They also asked for food recommendations, and while Gemini could help with restaurants, there was no follow-through on the Maps display to show you where to go.

The demonstration also shows what the Android Auto interface will look like after adding Gemini to the mix. A little indicator at the bottom of the screen will show that the AI is available to invoke. I imagine it’ll also be available through the integrated button in the steering wheel, just like Google Assistant is now.

Android Authority says the build of Android Auto it tested was spoofed on a smartphone rather than part of a car’s head unit. Nothing we saw in this video is finalized software, and Google hasn’t even officially declared that it’s here to stay. But still, I’m worried.

It’s been such a frustrating reality driving a car in this new age. I’ve already complained about the dongle in my old car, which is almost useless. I started receiving pop-up messages about Google Assistant’s impending deprecation immediately after writing that story. The Android Auto experience with the software built into the dash of my Subaru isn’t a walk in the park, either. It works most of the time for things like dictating messages and getting a clear direction to a specific place, but I still struggle to get it to point out gas stations along my route home versus behind me, where I came from. If I ever want to do anything that specific with a voice command in the car, I typically pull over to handle it instead. It’s too frustrating and distracting to use it the way it is now.

Gemini will hopefully be genuinely conversational. I want to be able to ask it to find me a gas station on my route that doesn’t require a significant detour off the main highway. I want to be able to say those exact words and have it understand my wants and needs.

I’ve been poking around with the chatbot recently, mainly in the browser, to get acquainted with this new tool I’ll inevitably have to adopt as an Android user. I noticed that if I change how I ask a question, or keep it super specific to the situation, I can get Gemini to catch on to where I need it to go and receive a helpful response. I wonder if the Android Authority demonstration asked the right questions. For instance, I wouldn’t have asked it to do the trip planning from inside the car. I’d ask instead for something more specific based on location, like, “Help me find the McDonald’s closest to a gas station along my route.” That’s the kind of thing I’d ask in the car and hope to get a computer’s help with. And that’s the specific, location-based usage case that would get me to accept that cars don’t have as many buttons as they used to. I hope that’s what Google has planned for Gemini in the car, whenever it launches.