I tried the Android Auto redesign and more Gemini is good here [Video]

Ahead of IO, Google announced a number of key changes to Android Auto and the Built-in in-car entertainment system that powers many modern vehicles. Here’s what it is like to use.

While often quietly updated and enhanced, Android Auto has become a very key component in Google’s arsenal over the past few years. Last year, we saw just what a Gemini-powered experience would look like, but this year, the company is driving forward with a more cohesive experience that finally adds some of the things lots of us have complained about for years.

For more video content, subscribe to 9to5Google on YouTube.

Fixing years of the same complaints

The good news is that Google has listened. Android Auto, in particular, needed some of the changes announced at the Android Show so desperately that it felt like they’d never arrive.

Finally adding YouTube feels like one of the easiest wins for in-car entertainment. The guardrails are sensible, and the ability to effectively listen to a YouTube video while driving is a huge boon for long-form content lovers. Podcast fans will be quietly rejoicing I’m sure. That said, it feels so long overdue when it has been possible with Google Built-in (formerly Android Automotive) for a while already.

Watching a bit of YouTube while parked is something I’m looking forward to very soon.

Adding in more detail to Google Maps with the Immersive View experience seems great on face value. I can certainly see it helping in areas where you might not be intimately familiar or have never visited previously. The 3D elements help you instantly position yourself without being too overwhelming. Just a few visual clues would have been enough, but this is an excellent implementation over the sometimes abstract shapes used currently. I can see myself getting lost less frequently in the future.

I am intrigued to see how this will work on every vehicle. Not all in-car displays are huge, almost borderless panels. The extra detail might be lost if your car has a small head unit that lacks visual clarity to see the intricacies.

That criticism aside, Immersive View appears to have reachable controls close to the driver. I liked this change as it makes micro adjustments quick, simple, and the focus isn’t taken away from controlling your vehicle.

However, a personal gripe is that every demo or hands-on session I’ve had with Android Auto or Google Built-in has been in a vehicle with a landscape head unit. We need to see more variation where possible, as lots of automakers like portrait-oriented displays – if you own a Volvo, you know what I’m complaining about here.

Touching in the differences with Google Built-in for a second, the ability to tie directly into in-car cameras and use those for even more precise navigation is something that feels right out of a Waymo. We were told that recordings are just a 30-second rolling snapshot, so that no files are kept locally.

The demo showcased asking for information about a landmark visible from the driver’s seat. While Gemini can be a private tour guide in this instance, the safety implications feel more valid. I could see myself using a feature like this to ask about road rules, get clarification on route closures, or even toll road terms. Real-time visual search without the associated risks of using your phone is still where I am sold on this kind of integration. Gemini is also able to reference car controls, settings, and even various dashboard information points.

I do worry about the increased power consumption and processing requirements. As impressive as some demos can be, losing data connection while accessing any cloud process kills the experience instantly. I can attest to this. At least I don’t need to bother my passengers when things go awry.

The hows and whys of in-car UIs

One of the few design changes I’m not really sure is all that great is the usage of widgets in the dashboard. Sure, it’s feature parity with your phone, but is this something that is really useful when you’re driving around? Why do I need to see a Google Photos widget in my car? Google Home controls make more sense, as do music player controls, but again, are they necessary?

Google is going deep on customizable widgets, and this is probably more useful for drivers despite my protestations. Sadly, we only saw yet another weather/condition widget from this demo. I guess this is the extent of what most people are doing in their cars during the testing phase.

It had me thinking, though, and my initial takeaway is that I’d prefer a customizable main homescreen view with integrated custom widgets. To me, this is probably the best way Google could have added the feature in vehicles rather than stashing these away out of sight until you summon them.

I find that the current default weather ticker is a little information-light. I’d like to have more details on the upcoming weather conditions in my area, or, going one step further, if I’m navigating to a different destination, it would be great to have Android Auto show the weather at that new location.

Maybe some safety concerns have limited what we were shown, but there is potential here if it can be integrated a little better.

I know a lot of people will also love that you can customize the Android Auto in-app wallpaper, but that felt so inconsequential that I don’t think it was even mentioned during my demo. It’s a part of the system that I try to avoid in all honesty, but it’s still going to be nice for those who care.

Widget and very minor wallpaper complaints to one side; Magic Cue feels more at home here than on my phone. The demo we saw had the data ready to go, but I’m intrigued as to how it’ll work in the real world. The less time I am interacting with the already massive display at the center console of my Volvo, the safer I feel. Magic Cue definitely feels like a sleeper addition.

Sadly, on phones, it has rarely manifested in my day-to-day life. Maybe with Gemini Intelligence set to come to Pixel phones later this summer, it’ll provide the backbone for a better experience. The Google Assistant was awful at voice dictation, whereas Gemini excels at it. Magic Cue being able to just take the lead and have responses ready based on my phone context seems like a killer feature. I just wish that we had more broad examples of it working than just asking for dates and times of an event but it is easy to demonstrate this working I guess.

A future of AI in cars

Hand on heart, I have steadily come around to AI integration – provided it is done correctly. The downside to that is that we simply do not know unless someone tries it first. Over the past 12 months, I have found myself more inclined to try various Gemini-powered features. Some are awful, some are genuinely useful.

As someone who actively avoids driving, I was very excited by Gemini in my vehicle. It needs a little work, but I think this is arguably the best place to integrate a contextually aware voice assistant. Mostly because automakers have decided that tactile buttons are just surplus to requirements. I hate fumbling with a touchscreen while driving. It’s distracting as it is potentially dangerous.

Gemini in vehicles is somewhat a solution to a problem created by automakers.

In my mind, Gemini summaries for travel should be a greater component of Android Auto and Google Built-in. For instance, I’m about to travel a few hours away. The weather is rainy in my location, but sunny at my destination. This data is available, but having Gemini remind me that an umbrella or a jacket might not be needed is something that could be useful. Least not sunscreen, or maybe a pre-drive checklist option would be good.

I’m sure, like me, you go into autopilot when in the car. I forgot so many important things because I’m focusing on the major task of driving, so proactive nudges from Gemini could be a truly powerful feature separator. Tie that into my Google account, which has intimate knowledge of my movements, and it’s a recipe for something that no other in-car entertainment system could potentially offer.

Ordering food from the driver’s seat is something that feels alien to my European mind. I have never felt the need to order something on such a whim. Google insists on hammering home that AI is ready to do that, but I just don’t think I trust it with the entire ordering process. Drive-thrus literally exist for this exact scenario. I think I’ll stick to that admittedly rare instance to sate my dining desires while driving.

As we saw with the big Gemini Live addition last year, we still have a potential age to wait before all this comes to our vehicles. I’m hopeful the demos are more accurate than last time, too, as Gemini’s rollout was a bit of a mess. That said, this feels like the perfect place to add even just a little more Gemini flavor.