AI Takes Spotlight At Google I/O 2025: Android XR, Gemini Updates, AI Mode In Search, Google Beam Announced

With a slew of AI-first breakthroughs, Google I/O 2025—the search giant’s flagship event—got underway. Gemini’s deeper integration took centre stage on Day 1, and Android XR was announced, along with Google Beam, AI Mode in Search, and more.

Here’s a look at all the major news and announcements from Google I/O.

Google showcased its Gemini-powered Android XR platform at Google I/O. With AI at its core, it will provide support for smart glasses and headsets, helping see and hear from surroundings and offering live translation, real-time messaging, and navigation. Gentle Monster and Warby Parker will integrate Android XR in their eyewear.

Project Astra’s (Google’s universal AI assistant) camera and screen-sharing features are now integrated into Gemini Live, which can be utilised for tasks ranging from interview preparation to training for marathons. While Android users have access to this feature, iOS users will begin to receive it now.

Google is launching a new AI Mode for an end-to-end AI Search experience. Users can ask lengthier and more complicated questions and follow-up questions in AI Mode, with advanced reasoning offering answers. Users can access this as a new tab directly within Search.

Google is making enhancements to Gemini 2.5 Pro by adding an improved reasoning mode called Deep Think. It leverages latest research in thinking and reasoning, including parallel thinking techniques, Google said. Also, the new Gemini 2.5 Flash, launching in June, improves reasoning, multimodality, code, and long context.

Google unveiled a new AI-first platform for video communications: Google Beam. Using six cameras and AI, Beam's video model combines 2D video feeds into a realistic 3D experience that is displayed on a 3D lightfield. The company said that later this year, the first Google Beam gadgets will be made available in partnership with HP.

Speech translation technology is coming to Google Meet to assist Workspace business users. It can mimic the speaker’s tone, voice, and even facial expressions in almost real time, Google said. Google AI Pro and Ultra subscribers will get translation in English and Spanish in beta, with other languages to follow.

Google is adding agentic features to the Gemini app, Chrome, and Search. The Gemini app’s new Agent Mode will enable users to do things like finding apartment listings. Subscribers will be able to access an experimental version of Agent Mode via the Gemini app. 

Google announced new customised Smart Replies feature in Gmail that can look through your Google Drive files and previous emails to provide a precise response. It will even mimic your usual greeting and capture your tone, manner, and preferred word selections. 

Deep Research will now allow users to upload files and connect to Google Drive and Gmail for generating custom research reports even more easily. Its integration with Canvas will enable creation of infographics, quizzes, and podcasts in multiple languages. 

Google presented Veo 3, its advanced video model featuring native audio production. Also, Imagen 4, the latest picture generation model, was announced. Both are available in the Gemini app.

The new tool will allow filmmakers to expand a brief clip into a longer scene and make cinematic clips.