You can access the original blogs from here.
Google used its I/O 2026 developer conference to signal a major shift from AI chatbots to so-called “agentic” AI systems, tools that can act on a user’s behalf across apps, devices, and the web. The company announced new Gemini models, AI-powered search upgrades, autonomous agents, shopping tools, developer platforms, and AI hardware, while also integrating its infrastructure and subscription ecosystem more deeply into everyday consumer and enterprise workflows.
Gemini Omni and Gemini 3.5 take centre stage: Google announced Gemini Omni, a new multimodal AI model that it says can eventually generate “any output from any input.” The first version, Gemini Omni Flash, can create and edit videos using text, images, audio, and video references. Google said the model combines Gemini’s reasoning with its media-generation systems to improve its understanding of physics and motion, resulting in more realistic outputs. The tool is being rolled out through the Gemini app, Google Flow, and YouTube Shorts.
Google also launched Gemini 3.5 Flash, the first model in what the company described as a new family focused on “frontier intelligence with action.” The company said the model is designed for coding, AI agents, and long-running tasks.
AI Search becomes more conversational and proactive: Search was another major focus of the event. Google said AI Mode has crossed 1 billion monthly users, while AI Overviews now reach 2.5 billion users every month. The company announced a redesigned AI-first Search interface that can handle text, images, videos, files and browser tabs. Google is also introducing ‘information agents,’ AI systems that continuously monitor topics, news, shopping trends, and finance updates in the background and send users summarized recommendations.
Google is also turning Search into a more interactive product. Powered by its Antigravity platform and Gemini 3.5 Flash, Search will now generate custom interfaces, visualisations, dashboards and mini-apps depending on the query. Some of these features will initially be limited to paid subscribers.
Gemini Spark and the push toward autonomous AI assistants: Another major announcement was Gemini Spark, a 24/7 AI agent that runs continuously in the background and can complete long-running tasks across devices and apps. Google said Spark works even when a laptop or phone is turned off because it runs on cloud-based virtual machines. The company plans to integrate it with third-party tools through the MCP protocol.
The Gemini app is becoming more proactive. Google introduced “Daily Brief,” which pulls information from Gmail, Google Calendar and Google Tasks to create a personalised morning summary. The company also redesigned the Gemini interface with a new system called “Neural Expressive,” which features interactive layouts, animations and voice-first controls.
New tools for creators and developers: For creators, Google announced Google Pics, an AI-powered image editing and design tool built on its Nano Banana image model. The company also upgraded Google Flow, its AI creative platform, with AI agents that can brainstorm, edit assets, organise projects, and create custom creative tools using natural language prompts.
Google’s developer announcements focused on Antigravity, its new agent-focused development platform. The company launched Antigravity 2.0, a desktop app that lets users manage multiple AI agents simultaneously. It also introduced Antigravity CLI, SDK support, Managed Agents in the Gemini API, and new Android app-building tools in Google AI Studio. Google said developers can now create Android apps directly from prompts and publish them to Google Play test tracks from within AI Studio.
The company also pushed deeper into AI subscriptions. Google launched a new $100-per-month AI Ultra plan aimed at developers and advanced users. The plan includes higher usage limits, expanded storage, and premium AI features.
AI glasses and scientific research tools: Hardware and XR also featured prominently. Google said its first AI-powered audio glasses, developed in partnership with Warby Parker, Samsung and Gentle Monster, will launch later this year. The glasses will support voice assistance, messaging, navigation, and real-time information overlays.
Google also announced new scientific AI tools under “Gemini for Science,” including systems for hypothesis generation, literature analysis and code-based scientific simulations. The company said the tools are designed to accelerate scientific research workflows
Sundar Pichai pitches the “agentic Gemini era”: Alongside product launches, Google repeatedly highlighted the scale of its AI operations. CEO Sundar Pichai said Google now processes “over 3.2 quadrillion” tokens per month across its products, up from 480 trillion a year ago. He said more than 8.5 million developers build with Gemini every month, and Google’s APIs process “roughly 19 billion tokens per minute.”
Pichai framed the event as a transition toward AI systems that actively perform tasks rather than simply answer questions. “We’re firmly in our agentic Gemini era,” he said.
At the same time, the event reflected Google’s attempt to defend and expand its position against growing competition from companies such as OpenAI and others racing to build AI agents, search products, and multimodal systems. Several announcements, including autonomous agents, AI-powered commerce, coding tools, and multimodal video generation, targeted areas where rivals have recently gained traction.
Google pushes transparency tools amid deepfake concerns: Google also tried to address concerns around AI-generated content and deepfakes. The company expanded its SynthID watermarking system and announced that OpenAI, Kakao and ElevenLabs will adopt the technology for some AI-generated content. Google said Search and Chrome will soon support verification tools that determine whether media was AI-generated or edited.
Throughout the event, Google positioned AI not as a separate product, but as infrastructure that will increasingly be integrated into Search, Android, Workspace, Chrome, YouTube and shopping. But many of the headline features, including agents, advanced search tools and creative systems, remain limited to testers, U.S. users or paid subscribers, with broader rollouts planned over the coming months.
Read more:
For You
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。