Gemini’s new app design makes me worried for the future of Android UI

It’s been a month since Google started rolling out its “Neural Expressive” design in Gemini across its mobile apps and the browser. Google also revealed that this new Expressive design powers Gemini Intelligence and will be pivotal in the way you interact with Android’s smart new experiences (such as integrations with food delivery apps) over the coming months.

But ever since I got the design and started using it, I can’t shake the feeling that it’s one step forward and several steps back for the Gemini app and, by extension, the future of Android’s entire design language.

Neural Expressive is a step back in usability

One of the most obvious changes in the new Gemini app, and the most recognizable aspect of Neural Expressive, is the new, thinner typography used throughout. At first, I thought it was a new font, but it’s the same Roboto I’ve been using on Android for years, just thinned out considerably. If you want to see it for yourself, check Google Fonts’ Roboto Flex page and drag the “Weight” bar down. The further you go, the closer you get to the Gemini app’s current typography.

Google intends for this to be a living font that reduces visual density, adapts, and scales with the entire interface. It’s basically made for a dynamic app design that changes depending on what the underlying smarts and AI require. The font gets heavier for more important bits, lighter for others. But the one clear side effect is that the default Roboto Flex font looks a little too thin and is a bit tougher to read than previous versions.

On top of this, Google has messed up a lot of the Gemini app’s obvious usability features. Now, the model selector is not next to the text box but at the top. I can understand that. What I don’t understand is mixing attachments with the choice of what you can make with Gemini in the same pop-up. Every time I use the + menu, I have to stop for a second and think if I need the big Photos icon or the smaller Create image icon. It’s as if Google forgot there’s a difference between adding something to a prompt and completely changing what the prompt does. And this scares me the most, because it feels like good app UX is taking a back seat while the focus is on making the interface purer.

Nothing, however, angers me more than the decision to move the account picker to the bottom of the side panel. For years now, every Android app has had its account switcher in the top right and allowed a simple swipe gesture to quickly flip between accounts. I’ve gotten used to this gesture to switch between my personal, work, and family accounts in Gmail, Google Search, Gemini, Contacts, Drive, Keep, Maps, Wallet, NotebookLM, the Play Store, and many more. It’s an awesome way to separate my online presence from work things on my personal account, or vice versa.

The switch to the side panel in the Gemini app ruins that. I have to swipe to open the panel, then swipe to switch accounts, which refreshes the app and closes the side panel. To switch again, I have to repeat that. It’s essentially transformed a quick swipe across the avatar in the top right into two separate and time-consuming gestures that I have to repeat multiple times. I hate it. And if this is a sign of where the account switcher will be in future updates of other Google apps, I might even start a riot.

Google should focus on the good aspects of Neural Expressive

Look, Neural Expressive is not all bad. It has fluid animations, nicer colors in both light and dark modes, slightly improved haptic feedback, and, in the new Gemini app, a more responsive mic and a Live mode that’s more integral to the experience rather than siloed in its own world.

All of this is incredibly positive and has improved the ways I interact with Gemini in the app. I wish Google would focus on these improvements and on making things more intuitive and interesting to use for everyone, instead of moving elements that have been in the same spot for years and years (like the account switcher) or cramming parts that don’t fit together into the same menu.

I shouldn’t have to relearn how to use the Gemini app every couple of months. Yet somehow, every few months, Google completely changes how the app presents itself, its models, and its options. My job is to keep up with this stuff, and I barely can, so how do we expect normal users to do that?