Gemini Live is one of Google's best AI tools to date, but it often feels more like a party trick than something you'd want to use daily. I intermittently used Gemini Live's multimodal camera and screen sharing features in situations where the extra context seemed important.
However, the basic Gemini chatbot remained my go-to AI feature for casual questions and everyday help. That all changed following a major Gemini app update that added a completely overhauled Neural Expressive user interface. Everyone is talking about the colorful and minimal Gemini app design or Gemini Daily Brief, but the feature I can't stop using following the update is Gemini Live.
Gemini Live's old voice-first user interface is gone, and in its place is a new approach that focuses on AI-generated content. It displays spoken words, generated images, and more on the screen as you go about your Gemini Live chat. You can copy, share, or export Gemini Live's responses without ending the conversation. It's the most underrated upgrade to come out of Google I/O, and I can't stop using it.
What's new about the Gemini app's UI?
The old Gemini Live experience made sense for long-running voice conversations or questions that required real-time multimodal context. It wasn't ideal for quick chats or basic questions. That's because the outgoing Gemini Live user interface took up the entire screen and didn't provide much visual value. There was a waveform graphic that moved as Gemini spoke, but you couldn't see or act upon anything it said.
The video experience was slightly better, as it displayed a viewfinder for the video stream being shared with Gemini. I've used Gemini Live's video streaming feature for everything from improving my plant health to rearranging my apartment layout. Still, these visuals were about the content you're sharing with Gemini, not the other way around.
Google's new Neural Expressive design language for Gemini Live condenses that waveform animation into a tiny pill. All the extra space is now used to highlight Gemini's responses. Like before, you can speak casually with Gemini Live and hear spoken responses uttered in real time. The difference now is that these responses will also appear on the screen, so you can read or listen to what Gemini Live has to say.
How I multitask using Gemini Live's new interface
The newfound focus on visual content enables new use cases for Gemini Live. You can ask it to generate images with Nano Banana 2, for example, and view the output immediately. The conversation remains active, allowing for follow-up tweaks and workshopping. This is something the old Gemini Live could never do.
To try it out, tap the waveform icon in the new Neural Expressive pill on the Gemini app home screen. The Gemini Live experience will begin, but you'll be kept on a tweaked version of the homepage.
Surrounding the Gemini Live animation pill, you'll see buttons for live video streaming, screen sharing, microphone mute, and exit. Get started by asking Gemini Live a question and watch the screen fill with the chatbot's answers.
This is where things get really useful. You can tap to interrupt Gemini Live, and when you do, a plain-text transcript of what the chatbot said appears — it's just like the one you'd see using the text-based Gemini version. This opens up a slew of controls and features. I can like, dislike, refresh, or copy the text of a Gemini Live response while the conversation is still going.
Previously, the old Gemini Live experience required users to end the live chat before the transcript and text controls would appear.
A response from Gemini Live can be exported to Docs, moved to a draft in Gmail, or used to branch a new chat. With the copy function, you can easily move a Gemini Live response to any other app without ending the conversation.
Gemini Live also works straight from your home screen. You can activate Live by holding down the power button or swiping up from the bottom-right corner, and tapping the Gemini Live waveform icon. The compact user interface makes even more sense for multitasking as you use your device.
As you continue to use your phone, the Gemini Live interface collapses into a tiny orb, leaving plenty of space for your content. All the while, you can still chat with Live. Tapping the circular overlay brings the buttons back, so you can share your screen, stream your camera feed, or mute your microphone. You can also hang up or mute Gemini Live from the live notification in the notification shade.
Gemini Live might be better than basic Gemini now
If I'm being honest, the main Gemini chat experience took a step backward with the Neural Expressive overhaul.
Things are harder to find, like how tools and attachments are strangely combined under a unified menu. There are no longer suggestion chips to spark ideas or help users experiment with new Gemini features. While the basic Gemini chat screen arguably became less functional, the Gemini Live experience only got more functional.
The updated version of Gemini Live lets you see text, images, and more on your screen as the voice AI helper speaks. I get almost all the benefits of using the chat experience without having to type or send off multiple queries. Now that Gemini Live is this useful, I don't see myself using anything else.
The best for Gemini Live
Google Pixel phones are made for multimodal Gemini Live, and the Neural Expressive interface fits right in with Google's Material 3 Expressive design language. The Google Pixel 10 also has a Tensor G5 chip, a 6.3-inch display, and a triple-camera system.
