Google I/O 2026 kicked off on May 19 with a keynote presentation that once again placed artificial intelligence at the center of the company's vision. Over the course of the event, Google announced a new family of Gemini models, a redesigned Gemini app, an always-on AI agent called Gemini Spark, and significant updates to Search, Gmail, Wear OS, and Android XR. Here is a detailed look at all the major announcements.

Gemini Gets a Redesign and a New Model

Google introduced a new design language for the Gemini app called “Neural Expressive,” featuring fluid animations, vibrant colors, new typography, and haptic feedback. The app now prominently displays a large “Ask Gemini” bar, with a subtle blue gradient at the bottom. Alongside the redesign, Google unveiled the Gemini 3.5 Flash model, which promises to be more intelligent and faster. The company also announced Gemini 3.5 Pro, targeting more complex reasoning tasks. These models are expected to power a range of Google services and third-party integrations.

Gemini Omni: A Multimodal Content Creation Model

A major new AI family called Gemini Omni was introduced, with the first model being Omni Flash. This model can generate videos from text, photos, videos, and audio inputs. Google envisions Omni as a tool that can “create anything from any input.” Omni Flash is positioned as a video counterpart to the popular Nano Banana image generation model, which has already been used to create over 50 billion images. Users will be able to insert their likeness into videos, a feature that builds on existing image insertion capabilities.

Gemini Spark: An Always-On AI Agent

Google launched Gemini Spark, its version of an always-on AI agent similar to OpenClaw. Powered by Gemini 3.5 Flash, Spark runs 24/7 on virtual machines in Google Cloud. It can write emails, create study guides, monitor credit card statements for hidden fees, and connect to Workspace apps like Gmail, Docs, Sheets, and Slides. It also integrates with third-party services like Canva, OpenTable, and Instacart via the Model Context Protocol (MCP).

Google Search Gets Its Biggest Changes Ever

Google Search is undergoing a major overhaul. The search box now dynamically expands for longer queries and offers AI-powered autocomplete suggestions that go beyond traditional predictions. AI Overviews, the AI-generated summaries, are now more reliably shown for natural language questions. Users can also attach documents, photos, videos, and Chrome tabs to the search box to jump directly into AI Mode, which functions like a chatbot. The new updates are powered by Gemini 3.5 Flash.

Gmail Live: Voice Mode for Your Inbox

Gmail is getting a new AI-powered voice feature called Gmail Live. Similar to Gemini Live, it allows users to ask questions and get answers from their inbox by simply speaking. In a demo, a user asked about events at a child's school and an upcoming trip, with Gmail Live instantly pulling relevant details like dates and locations from emails.

Android XR Glasses: Project Aura Progress

Google showed off updates to its Project Aura smart glasses, developed in collaboration with Xreal. The glasses now come with a redesigned compute puck that includes a fingerprint scanner and a lanyard for neck wear. The glasses themselves are a dark pair of sunglasses that sit between a full headset and lighter mixed-reality glasses. Google also teased partnerships with Warby Parker and Gentle Monster for future designs. The Android XR platform is expected to launch later this year.

Wear OS 7: Live Updates and AI Features

Wear OS 7 was announced with several new features. It will support iPhone-style Live Updates, which can appear on the watch or connected smartphone, tracking deliveries, sports scores, and automated AI tasks. The platform is also getting “Wear Widgets,” which align with Android's 2x1 and 2x2 widget formats. Additionally, select watches will receive Gemini Intelligence later this year, bringing personalized and proactive AI assistance.

Google Pics: AI Image Editing for Workspace

Google launched a new Workspace app called Pics, designed to simplify AI image editing. Instead of rewriting entire prompts, users can click on parts of an image and leave notes to make targeted changes. The app is powered by Gemini and the new Nano Banana 2 image model. In a demo, a user modified a party invite by clicking on a cat and requesting a dog, though the result still resembled a cat.

AI Detection and Labeling: SynthID and C2PA

Google expanded its AI detection capabilities. SynthID watermarks and C2PA content credentials are now more accessible through Chrome and Search. Users can check if an image was AI-generated by selecting it in Google Lens, Circle to Search, or AI Mode. Chrome will soon allow users to right-click and ask “Is this made with AI?” to see provenance information. OpenAI also announced it will adopt SynthID for its models.

Universal Cart: AI-Driven Shopping

Google unveiled a “Universal Cart” that works across different retailers and Google products like Gemini, YouTube, and Gmail. Users can add products to the cart from Search or Gemini conversations and check out through Google. The cart tracks prices, provides stock notifications, suggests discounts, and alerts shoppers to potential issues.

Google AI Studio Goes Mobile and Builds Android Apps

Google announced that its AI Studio vibe coding tool can now build native Android apps. Users can prompt an app idea, preview it in an embedded emulator, and install it on a connected phone. Pre-registration for an Android version of AI Studio is available on Google Play.

AI in the Car: Gemini and Volvo

Gemini will integrate with Volvo's external cameras in the upcoming EX60 SUV. The AI assistant can interpret parking signs, lane markings, and nearby landmarks by accessing the vehicle's cameras. This is possible thanks to Volvo's use of Google's embedded Android Automotive.

Cybersecurity Push: CodeMender

Google is expanding access to CodeMender, an AI agent for code security, allowing external experts to use its API. The tool flags and fixes vulnerabilities, positioning Google to compete with Anthropic's Mythos and other cybersecurity AI offerings.

Gemini for Science and the Singularity Mention

DeepMind CEO Demis Hassabis closed the keynote by introducing Gemini for Science, a set of tools for scientific research, potentially aimed at drug discovery. He suggested humanity may be at the “foothills of the singularity,” sparking discussion about AI's transformative impact.

Other announcements included a new Google Pics app, updates to Google TV with pointer remote support, and a reduction in the price of the Google AI Ultra plan from $249.99 to $100 per month, with a $200 tier that includes access to Project Genie world models.

Source: The Verge News