Google still hasn’t fixed my biggest problem with Gemini on Android Auto [Video]

Gemini for Android Auto has been divisive, to say the least, but for all of its pros and cons, there’s still one big problem that frustrates me on a frequent basis.

Gemini for Android Auto has been a marked improvement over the Assistant, though the bar was very low given Google had pretty much abandoned its prior voice assistant in its Gemini era. But outside of just fixing core functionality, Gemini Live turned out to be one of my favorite upgrades, as having a conversation while behind the wheel was a great way to brainstorm ideas and talk out problems in time that’s otherwise unproductive.

For all of the good things, though, I still have one problem with Gemini on Android Auto.

It just won’t shut up.

That is to say, Gemini continues speaking after I’ve already completed an interaction. Behind the wheel, voice interactions are useful because they’re all about lessening the amount of time I need to interact with the touchscreen and, in turn, take my eyes off of the road.

Commands such as “navigate to [location]” and “add a stop at [insert name here]” often trigger Gemini to bring up a list of options in Google Maps. Those options are shown on screen in Maps, while Gemini starts talking through the list so you can complete the entire interaction with voice commands. It’s a great combination, but the problem is that Gemini doesn’t stop talking when you choose to interact with the touchscreen instead. If I use Gemini to start the process – which is the only choice when your vehicle is in motion – the assistant will keep speaking even if I’ve already selected and confirmed my destination via Maps. “Which one would you like?” Gemini would say, sitting there waiting for a response despite the options already gone from the display and my route already in place.

Here’s what I’m talking about (“Hey Google” warning):

It’s that disconnect that frustrates me. Assistant on Android Auto never had this issue, with the touchscreen interactions ending the voice commands because, obviously, the user has already completed what they sought to do in the first place.

This has been a problem with Gemini for months now, since the very first rollout landed on my devices back in November of last year. And yet, Google hasn’t fixed it. Occasionally I’ll notice that Gemini understands I’ve interacted with the touchscreen instead of it – imagine that, a manual action in the age of AI agents – but it’s never consistent.

Google, please fix this. There are a ton of issues with Gemini on Android Auto, as users have rightfully been calling out, but this one is mind-bogglingly simple and staring you right in the face.

Follow Ben: Twitter/X, Threads, Bluesky, and Instagram

FTC: We use income earning auto affiliate links. More.