Google I/O 2026 kicked off on May 19th with a keynote packed full of artificial intelligence announcements. The company once again put Gemini at the center of its strategy, introducing new AI models, an always-on agent called Gemini Spark, a redesigned app, and ambitious plans for everything from search to smart glasses. The event also included updates on Android 17, Wear OS 7, Gmail, and Project Aura, Google's long-rumored augmented reality glasses.

Gemini 3.5: A New Family of Models

The most significant AI reveal was the Gemini 3.5 model family. Google introduced Gemini 3.5 Flash as a more intelligent and faster model for everyday tasks, and Gemini Omni Flash, a generative model that can create videos from text, images, audio, and video inputs. The company touts Omni as a step toward a model that can 'create anything from any input.' Omni Flash is positioned as the video equivalent of the popular Nano Banana image generator, which has already been used to produce over 50 billion images. Google also announced Gemini 3.5 Ultra, the most capable model, which powers the newly lowered subscription tiers: the Google AI Ultra plan drops from $249.99 to $100 per month, with a $200 option that adds access to Project Genie world models.

Gemini App Redesign: 'Neural Expressive'

The Gemini app itself is getting a major visual overhaul called 'Neural Expressive.' The design language features fluid animations, vibrant gradients, haptic feedback, and a large 'Ask Gemini' bar. Google says this makes the app feel more conversational and friendly. The update also brings deeper integration with Google services, allowing users to pull information from Gmail, Docs, and Calendar with natural language queries.

Always-On AI Agent: Gemini Spark

One of the boldest announcements was Gemini Spark, an always-on AI agent that runs 24/7 on virtual machines in Google Cloud. Powered by Gemini 3.5 Flash, Spark can write emails, create study guides, monitor credit card statements for hidden fees, and more. It connects to Workspace apps and third-party services like Canva, OpenTable, and Instacart via the Model Context Protocol (MCP). This positions Google as a direct competitor to Anthropic's OpenClaw platform, which also offers persistent AI assistants. Google is betting that users will trust an agent that can spend money, access personal data, and act autonomously.

Search Gets Its Biggest Changes Ever

Google Search is undergoing a transformation. The search box now dynamically expands for longer queries and offers AI-powered suggestions that go beyond autocomplete. AI Overviews, the AI-generated summaries at the top of results, are now powered by Gemini 3.5 Flash and appear more reliably when asking natural-language questions. Users can also attach documents, photos, videos, and Chrome tabs to the search box for a seamless transition into AI Mode, a chatbot-like experience. Google's vice president of product for Search, Robby Stein, said the goal is to make searching feel like a conversation.

Android XR Glasses: Project Aura Takes Shape

After years of speculation, Google finally showed off its Android XR smart glasses in detail. The Project Aura prototype, developed with Xreal, is a dark pair of sunglasses with a compute puck that can be clipped to a waistband or worn around the neck via a lanyard. The puck now has a fingerprint sensor for easy unlocking. Google also confirmed partnerships with Warby Parker and Gentle Monster for consumer models launching later this year. The glasses support hands-free notifications, live translation, and AI-powered contextual assistance. This marks Google's most serious push into augmented reality since Google Glass.

Wear OS 7: Live Updates and Smart Tiles

Wear OS 7, the next major smartwatch platform update, brings iPhone-style Live Updates to the wrist. Users can track deliveries, sports scores, and AI task progress directly from their watch. Tiles are being replaced by 'Wear Widgets,' which mirror Android's 2x1 and 2x2 widget formats for glanceable information. Select watches will also gain 'Gemini Intelligence' later this year, offering personalized, proactive AI features. Google is clearly aiming to make wearables a more integrated part of its AI ecosystem.

Gmail Live: Talk to Your Inbox

Gmail is getting a voice mode called Gmail Live. Much like Gemini Live, users can tap an icon and ask questions about their emails—such as event details, travel plans, or receipts—and the AI will respond verbally while pulling up relevant information on screen. During a demo, a Google employee asked about her child's school event and a trip to Detroit, and Gmail Live instantly surfaced dates, locations, and confirmations from her inbox.

AI Detection and Deepfake Labeling

Google also expanded its efforts to combat AI-generated misinformation. SynthID watermarks and C2PA content credentials are now integrated into Chrome and Search. Users can select or circle an image and ask 'Is this made with AI?' and the system will check for provenance markers. OpenAI simultaneously announced it would adopt SynthID watermarks for its own outputs, creating a cross-industry standard.

Other Notable Announcements

Google launched a new app called Pics for AI image editing within Workspace, allowing users to click on parts of an image and leave notes for edits. AI Studio can now build native Android apps through vibe coding. Android 17 was previewed with AI-generated widgets and deeper Gemini integration. Volvo announced that Gemini will use external cameras in its EX60 SUV to interpret parking signs. Google also revealed a 'Universal Cart' that lets users add products from different retailers and check out through Google—a direct challenge to Amazon.

Inside Google's Beam Lab, reporters met Sophie, a lifesize AI agent that can see and talk. The demo hinted at Google's ambition to create physical AI companions. Demis Hassabis, CEO of Google DeepMind, closed the keynote by saying we might be in the 'foothills of the singularity,' announcing Gemini for Science tools to reimagine drug discovery. The event made clear that Google is all in on AI, trusting users to share their personal data for a more proactive and predictive digital future.

Source: The Verge News