Google I/O 2025: Gemini AI Expands with Beam, Flow, Jules, and Android XR Innovations

Update: 2025-05-21 13:15 IST
Google I/O 2025: Gemini AI Expands with Beam, Flow, Jules, and Android XR Innovations
  • whatsapp icon

At the highly anticipated Google I/O 2025 conference held on May 20, Google showcased a series of transformative AI innovations centered round Gemini, the company’s flagship family of large language models. From animmersive 3D communication platform and autonomous coding assistants to afilmmaking tool powered by AI, the event offered a sweeping vision of howartificial intelligence will shape the future of Google's products andservices.

“The opportunity with AI is truly as big as it gets,”declared Sundar Pichai, CEO of Google, setting the tone for a developerconference that leaned heavily into AI-first experiences.

Here’s a comprehensive breakdown of the most significant announcements from Google I/O 2025:

Google Beam: A Leap in 3D Video Communication

In a major step forward for remote collaboration, Google introduced Beam, the next evolution of its earlier Project Starline. Beam use ssix cameras combined with an AI-powered volumetric video model to convert 2Dvideo streams into real-time, immersive 3D experiences.

The technology offers millimetre-accurate head tracking and renders at a smooth 60 frames per second, delivering a lifelike, in-person meeting feel—even across long distances. Beam is built on Google Cloud, promising enterprise-level reliability, and is being rolled out in partnership with HP and Zoom. Live demonstrations will be showcased at Info Comm, with availability for select enterprise customers later this year.

Gemini App Expands with AI Live, Veo 3, and Imagen 4

The Gemini app, now with over 400 million monthly active users, has been positioned as a universal AI assistant. One of its standout features is Gemini Live, which integrates capabilities from Project Astra, including camera and screen-sharing features. Users can point their smartphones at objects or share their screens for real-time AI-powered support, useful for everything from prepping for interviews to marathon training.

New tools like Imagen 4 and Veo 3 have also been added. Imagen 4 enables high-quality image generation, while Veo 3 allows for vide creation with native sound effects and dialogue.

Gemini is also integrated into Chrome, enabling users to askquestions directly while browsing.

To expand access, Google has introduced two new subscriptionplans:

• Google AIPro ($19.99/month or approx. ₹1700), offering tools like Flow and Notebook LM with higher usage limits.

• Google AIUltra ($249.99/month or approx. ₹21,400, US only), includes experimental features like Agent Mode and access to the most advanced Gemini models. Students in select countries, including the US, UK, Brazil, Indonesia, and Japan, can access AI Pro free for one year.

AI Mode in Google Search Gets Smarter

Google Search is being revamped with AI Mode, powered by the custom Gemini 2.5 model. It is now rolling out in the US without the need forLabs sign-ups. This upgraded search experience handles multimodal queries, supports follow-up questions, and offers links to trusted web sources.

Other AI-powered enhancements include:

• Deep Searchfor in-depth, contextual answers.

• SearchLive, a visual search tool that uses the phone’s camera to identify objects and answer real-time queries.

According to Google, AI Overviews now serve 1.5 billion users across 200 countries and have led to a 10% increase in search queries incountries like the US and India, making it one of Google’s most successful search features in a decade.

Project Mariner & Agent Mode: Teaching AI to Work for You

Project Mariner, Google’s prototype AI agent system, has received major upgrades. It can now handle up to 10 simultaneous tasks, such as booking tickets, conducting research, or shopping online. Mariner features a “teach andrepeat” system that enables it to learn tasks after just a single demonstration.

These features will first roll out to Google AI Ultra subscribers in the US, with developer access arriving through the Gemini APIlater this summer. Companies like Automation Anywhere and Ui Path are already experimenting with Mariner to power business automation.

Flow: AI-Powered Filmmaking for All Creators

Google introduced Flow, a filmmaking tool designed for both beginners and professionals. Powered by Gemini, Imagen, and Veo, Flow offers:

• Cameracontrols for adjusting motion and camera angles.

• Scenebuilder for editing or extending video shots.

• Asset management tools to keep prompts and creative elements organised.

The platform also includes Flow TV, a curated showcase of AI-generated videos, complete with visible prompts to inspire creativity.

With Veo 3 integrated, Flow supports dialogue, sound effects, and ambient audio, enabling creators to produce professional-grade content using only AI.

Jules: A Smarter, Autonomous Coding Assistant

First introduced via Google Labs in December, Jules has now entered public beta. It is a powerful, asynchronous coding assistant available anywhere Gemini operates.

Unlike traditional code-completion tools, Jules can:

• Write andtest code.

• Fix bugs and update dependencies.

• Generate audio-based changelogs.

Jules integrates securely with repositories on Google Cloud virtual machines and does not train on private code, preserving privacy. It provides detailed plans and reports after task completion, leveraging the Gemini 2.5 Pro model to handle complex, multi-file coding challenges.

Stitch: Bridging Design and Development

Stitch, another Labs experiment, aims to streamline appdevelopment. It uses Gemini 2.5 Pro’s multimodal abilities to transform text prompts and images into functional UI designs and frontend code.

Stitch features include:

• Interactive chat support.

• Theme selectors.

• Integration with Figma, enhancing collaboration between designers and developers.

By simplifying the workflow between design and engineering, Stitch helps teams build more intuitive, efficient applications.

Android XR: Google’s Step into Spatial Computing

Google also revealed major developments in Android XR, its spatial computing platform designed for headsets and smart glasses. Built incollaboration with Samsung and Qualcomm, Android XR powers upcoming wearables like Samsung’s Project Moohan headset, expected to launch later this year.

Android XR glasses come equipped with:

• Cameras,microphones, and optional in-lens displays.

• Seamless Gemini integration for real-time assistance, including messaging, navigation, and live language translation.

The goal is to make AI truly context-aware—understanding what users see and hear to offer intelligent, hands-free help.

Gemini 2.5 Series: Pro and Flash Upgrades

The Gemini 2.5 series, which includes Pro and Flash models, has received significant updates. These include:

• Nativeaudio output for natural conversations.

• Enhanced security features.

• Integration with Mariner for desktop tasks.

Gemini 2.5 Pro now features Deep Think, an experimental mode for tackling complex math and programming problems. Meanwhile, 2.5 Flash is optimized for speed and is the default model in the Gemini app.

Both versions will be available on Google AI Studio and Vertex AI by early June, offering powerful tools for both developers and enterprise users.

Gemini in Workspace: Smarter Tools for Productivity

Finally, Google announced Gemini-powered enhancements across its Workspace suite, including Gmail, Docs, Meet, and Vids. These updates allow users to:

• Respondto emails more quickly.

• Translatemessages across languages.

• Createvideos and write documents with relevant sources.

The goal is to boost productivity by making AI an integral part of daily work.

From powerful developer tools and intelligent agents to creative filmmaking and immersive communication, Google I/O 2025 made it clear: the Gemini AI ecosystem is set to transform how we interact with technology—at work, at home, and everywhere in between.

Tags:    

Similar News