Google I/O 2026: AI Agents Power New Search, Gemini Updates

MOUNTAIN VIEW, Calif. – Google's 2026 I/O developer conference concluded its keynote address on May 20, 2026, unveiling a broad suite of new agentic artificial intelligence features and advancements. The company demonstrated upcoming innovations for its Android-powered smart glasses and prominently featured its ongoing AI initiatives, which are set to become central to its major services. Google CEO Sundar Pichai emphasized the company's period of "hyper progress" in AI, noting the user demand for tangible product value. This year's event focused on embedding AI agents into Search, Gmail, YouTube, Docs, and Chrome.

The cornerstone of the announcements was the overhaul of Google Search, now featuring an "intelligent search box" powered by AI agents. This new experience, rolling out immediately, aims to provide more contextual answers and generate custom images or short video clips to explain complex topics. For instance, a search query about black holes could yield an AI-generated video illustrating the concept directly within the search results. Additionally, Google introduced Generative UI, a feature that dynamically creates tailored layouts for search results, adapting to display videos, images, or news articles based on relevance. This feature is slated for release across all Google searches this summer. Search agents are also expanding their reach, integrating with Google Maps through the existing Ask Maps functionality for chatbot-like queries.

Gemini Models and App Enhancements

Google announced two significant updates to its AI models: Gemini 3.5, now officially released, and Gemini 3.5 Flash, a more streamlined and cost-effective variant. Both models are immediately accessible through Google Search and the core Gemini app. The Gemini app itself received a visual refresh named Neural Expressive, featuring new colorful backgrounds, an updated typeface, and enhanced animations for live voice chats. The app also introduces expanded voice accent options, including regional dialects in multiple languages. A new feature called Daily Brief is set to launch, offering users a personalized morning digest of their day by compiling information from calendars, emails, and other data sources to prioritize upcoming plans. Google is also integrating Gemini-powered features into other services. Ask YouTube allows users to ask natural language questions, receiving answers compiled from relevant video segments, even snapping to specific timestamps. Google Docs is introducing voice editing capabilities through a feature called Docs Live, where users can dictate content, generate text, and pull citations from the web, transforming spoken ideas into structured documents. For power users, Google is introducing a new subscription tier, AI Ultra, priced at $100 per month, while also reducing the cost of the existing Gemini AI Ultra from $250 to $200 per month.

The company also unveiled Gemini Omni, an AI video generator designed for realistic video creation, capable of incorporating real footage and generating elaborate AI-driven effects. This tool allows users to modify selfie videos with different backgrounds, styles, or environments, enabling them to appear in diverse virtual settings. Demonstrated onstage was a feature where a user's video of walking through a sculpture could be transformed to make the structure appear as if made of bubbles. Users can also add personal images and videos to generate cinematic styles. While Omni initially focuses on video, capabilities for still images and text generation are planned for later. An introductory version, OmniFlash, is available today for Google AI+ Pro and Ultra subscribers. Gemini Spark was also introduced, positioned as a personal AI agent capable of assisting with daily tasks like writing emails, planning events, and managing long-term projects by accessing user files and learning personal schedules. Spark operates entirely on Google Cloud, enabling background processing without draining device battery. Initial integration is with Google software, with Chrome browser and third-party support expected later this summer.

Google is enhancing its shopping experience with an agentic-powered system. As users search for products, Google will display listings from various retailers, consolidating them within a universal shopping cart. This agent can manage wish lists, alert users to price changes, new product versions, or color options. Users can interact with Gemini within the cart to request more product details, add items, or find better deals. Purchases can be completed directly through Google's secure payment system within the cart, even for items from multiple stores. An alternative option allows users to complete purchases on the original retailer's websites. The company also highlighted updates to its Flow toolset, designed for creating videos, images, and music, adding capabilities to simplify the creation of promotional videos, invitations, and short films for users with creative projects in mind.