Ainsider
Posts
Gemini AI Features at Google I/O 2025

Gemini AI Features at Google I/O 2025

Google Gemini is becoming AI operating system? Summary of Google I/O 2025 with huge Gemini AI updates and new features

Piotr Macai
May 21, 2025

Introduction to Gemini AI at Google I/O 2025

Google I/O 2025, held on May 20, 2025, marked a pivotal moment for artificial intelligence, with Google unveiling significant enhancements to its Gemini AI model. This survey note explores the AI features, functions, and possibilities of Gemini AI, providing a comprehensive overview for researchers, developers, and enthusiasts. The analysis covers the transformation of Gemini into an AI operating system, key feature updates, subscription models, additional offerings, hardware innovations, and future implications.

Research suggests Gemini AI at Google I/O 2025 is a versatile AI operating system with new features.
It seems likely Gemini Live enhances real-time interactions with camera and app integrations.
The evidence leans toward Imagen 4 and Veo 3 improving image and video creation with advanced capabilities.
Gemini in Chrome and Agent Mode may streamline browsing and task automation, with some features coming soon.
Subscription plans like Google AI Pro and Ultra offer advanced access, with pricing at $19.99 and $249.99/month.

Analysis of Gemini AI Features at Google I/O 2025

Gemini as an AI Operating System

Google has positioned Gemini as an AI operating system, a foundational platform integrating various tasks and applications, akin to traditional operating systems like Windows or Android. This evolution suggests Gemini now handles complex, multi-modal interactions—text, voice, vision—while seamlessly integrating with Google services. This shift indicates a future where Gemini acts as a central hub for AI-driven productivity and creativity, potentially reshaping user-device interactions.

Key Features and Updates

The following table summarizes the core features announced at Google I/O 2025:

Feature	Description	Availability	Key Capabilities
Gemini Live	Real-time camera interaction, integrates with Maps, Calendar, Tasks, Keep	Free on Android & iOS	Point at objects, get instant info, enhance productivity
Imagen 4	High-res image generation with accurate text, supports aspect ratios	Free in Gemini app	Create posters, pitch decks, artistic designs
Veo 3	Video model with native sound, 4K, long-form, multiple characters	Part of Ultra plan, free trials	Generate films, educational videos, marketing content
Deep Research	Upload files, images, PDFs for AI-powered reports, future Drive/Gmail integration	Available now	Research assistant for academic, professional use
Canvas	Creative studio for code, quizzes, infographics, podcasts via text prompts	Powered by Gemini 2.5 Pro	Rapid content creation for educators, marketers
Gemini in Chrome	Browsing assistant for summaries, questions, soon task automation	Rolling out to subscribers, US	Enhance productivity, reduce tab-switching
Interactive Quizzes	Generate quizzes with feedback, personalized follow-ups	Available now	Personalized learning, self-assessment
2.5 Flash	Fast, lightweight Gemini model, now default	Free	Broad accessibility, improved performance
Agent Mode	Autonomous task completion (e.g., apartment booking), uses Project Mariner, MCP	Coming soon to subscribers	Delegate complex tasks, enhance personal assistance
Google Flow	Powerful new type of AI filmmaking and video editing tool that combines the best of Veo, Imagen and Gemini — built with and for creatives. Flow helps you maintain character and visual consistency from one clip to the next.	Rolling out to subscribers, US	Video editing tool

Detailed Feature Analysis

Gemini Live: This feature allows users to point their camera at objects for real-time assistance, integrating with Google apps for tasks like planning or shopping. Its free availability on Android and iOS suggests broad accessibility, with possibilities for hands-free, context-aware assistance in daily life.

See it, solve it: Gemini Live's camera and screen sharing is available for free and it’s rolling out on Android and iOS starting today.
#GoogleIO
— Google Gemini App (@GeminiApp)
6:16 PM • May 20, 2025

Imagen 4: Generates high-resolution images with accurate text, supporting aspect ratios like 16:9. Its free availability in the Gemini app democratizes design, enabling users to create professional visuals without expertise.

Now you can generate images with Imagen 3 directly in Google Gemini chat window 📸
Imagen 3 is Google the latest image generation model with major improvements:
- higher degree of photorealism,
- better instruction following,
- fewer distracting artifacts
than ever before.
— Macai (@piotrmacai)
7:29 AM • Oct 10, 2024

Veo 3: Supports 4K video, long-form content, and multiple characters with dialogue. Part of the Ultra plan, it’s a game-changer for filmmaking, with potential to reduce costs and time in content creation.

We’re also launching Veo 3, our state-of-the-art video generation model.
Veo 3 lets you generate videos with sound effects, background noises and even dialogue.
#GoogleIO
— Google Gemini App (@GeminiApp)
6:23 PM • May 20, 2025

Google Flow

Powerful new type of AI filmmaking and video editing tool that combines the best of Veo, Imagen and Gemini — built with and for creatives.

Flow helps you maintain character and visual consistency from one clip to the next.

Introducing Flow: a new type of AI filmmaking tool that combines the best of Veo, Imagen and Gemini — built with and for creatives.
Flow helps you maintain character and visual consistency from one clip to the next.
See how emerging filmmakers are using it 🎥
— Google (@Google)
6:35 PM • May 20, 2025

Powerede by Gemini 2.5 Flash: Fast and lightweight, now default, ensures broad accessibility, with native audio outputs across languages.

— Google Gemini App (@GeminiApp)
7:24 PM • May 20, 2025

Canvas: Powered by Gemini 2.5 Pro, enables rapid content creation, from code to podcasts, enhancing creative workflows for non-specialists.

Canvas is getting updates, too: Now, you’ll see a “Create” button in your chats that allows you to generate interactive content based on your chats with just one click–no prompting required.
You can even transform your Deep Research report into a custom web page, podcast or an
— Google Gemini App (@GeminiApp)
6:18 PM • May 20, 2025

AI Mode: Initially for subscribers in the U.S., enhances productivity by reducing tab-switching, with plans for task automation.

AI Mode is rolling out to everyone in the US. It’s a total reimagining of Search with more advanced reasoning so you can ask longer, complex queries.
AI Overviews are now used by 1.5B people a month, in 200+ countries and territories.
And Gemini 2.5 is coming to both this week.
— Sundar Pichai (@sundarpichai)
7:25 PM • May 20, 2025

Deep Research: Combines user data with public info for reports, with future Drive/Gmail integration. It’s a powerful research assistant, with applications in academia and market insights.

Starting today, you can upload your own files and images into Deep Research, one of our top requested features.
Soon, you’ll even be able to connect your Google Drive and Gmail so you can pull info from there, too.
#GoogleIO
— Google Gemini App (@GeminiApp)
6:17 PM • May 20, 2025

Agent Mode: Upcoming for subscribers, promises to delegate complex tasks like apartment booking, a step toward autonomous AI assistance.

Agent Mode in the @GeminiApp can help you get more done across the web - coming to subscribers soon.
Plus a new multi-tasking version of Project Mariner is now available to Google AI Ultra subscribers in the US, and computer use capabilities are coming to the Gemini API.
— Sundar Pichai (@sundarpichai)
7:27 PM • May 20, 2025

Subscription and Advanced Features
Google introduced Google AI Pro ($19.99/month) and Ultra ($249.99/month) plans, offering access to advanced tools like Flow and Veo 3. Agent Mode, coming soon, promises autonomous task completion like apartment booking, leveraging Project Mariner and MCP.

Additional AI Offerings and Integrations

Beyond core features, Gemini’s ecosystem includes:

Project Astra: Revamped with native audio, UI control, content retrieval, calling, shopping, enhancing interaction.
Project Beam: A 3D video platform with HP, using six-camera setups for lifelike avatars, launching later 2025.
Real-time Translations in Google Meet: Uses Project Starline for English/Spanish, expanding language support.
Jules: Asynchronous coding agent, public beta, for planning and modifying codebases.
Stitch: Google Labs product for design generation, integrable with Figma.
Try it on: AI-powered virtual try-on in Google Search Labs for shopping.
AI Mode in Google Search: Rolling out to U.S. users, offering direct answers like ChatGPT, with personalized results from Gmail.
Hardware Innovations: Hardware announcements include Samsung’s Project Moohan headset and AI smart glasses, set for later 2025 release, integrating with Gemini for immersive experiences, competing with Apple Vision Pro and Meta Ray Ban Glasses.

Conclusion

Google I/O 2025’s Gemini AI advancements, from real-time interactions to autonomous task completion, indicate a future where AI is seamlessly integrated into daily life. These features promise to enhance productivity, creativity, and global communication, with implications for education, entertainment, and e-commerce. As these technologies roll out, they may reshape industries, offering personalized, efficient, and immersive experiences for users worldwide.

Key Citations