Google's redesigned Gemini comes with a new interface and AI models

Google has announced several updates for the Gemini app at its I/O annual developer conference, including a design language called "Neural Expressive" that gives it a redesigned interface with  new typography, more fluid animations and haptic feedback. The company has also integrated Gemini Live into the main Gemini experience. 

Currently, you'll have to tap the Live button in order to have spoken conversations with the AI, but the redesign will allow you to easily switch between typing and talking to the chatbot. Google is also introducing regional dialects for Gemini and designing its responses with imagery, graphics and even narrated videos, so it doesn't respond with walls of texts. The redesigned chatbot is now rolling out worldwide on Android and iOS. 

A new AI agent called Daily Brief will also be available starting today. If you opt in, Daily Brief will gather information from all your connected apps in the background, such as Gmail and Calendar, to give a summary of your activities for the day, as well as any communication waiting for you. In addition to giving you a quick overview of your emails and upcoming events, it can prioritize the things you have to do for the day based on your goals and will suggest the next steps you can take. You can train the agent to be more in line with your thinking by giving its responses a thumbs up or down. Daily Briefing is rolling out today, starting in the US, as part of Google AI Plus, Pro and Ultra subscriptions. 

Google has also released a new model called Gemini Omni that combines text prompts, images and videos you upload to generate video output. You can change video backgrounds with a text prompt and apply effects, as well as built-in templates, to the output. If you want to insert yourself in videos, it can create an avatar the looks and sounds like you. The new model will be available today for all Google AI Plus, Pro and Ultra subscribers.

Another new agent is Gemini Spark, which the company describes as a "24/7 personal AI agent." It runs on Gemini 3.5 and is deeply integrated with Workspace apps like Gmail, Doc and Slides. It can also be linked other apps like Canva, OpenTable and Instacart. You can use it perform recurring tasks, such as poring over your credit card bills every month for hidden subscription fees. You can also create complete workflows from it. Spark can, for instance, create notes from information across emails and chats, write up a report on Google Docs and then draft an email from that information to create a new project. Spark is making its way to testers this week, before rolling out to Google AI Ultra beta subscribers in the US. 

Finally, Google is bringing Spark to the Gemini app for macOS this summer, along with new voice experiences. Even if you use filler words and expression while you speak, Gemini will be able to turn your speech into precise drafts that you can then use for your reports or emails.