← Back to articles
Google announces AI agents across Search, Gmail, YouTube, and new Gemini models at I/O 2026
WIRED AI · May 19, 2026
AI Summary
- •Google released Gemini 3.5 and Gemini 3.5 Flash today, along with a redesigned Gemini app featuring Neural Expressive styling, new voice accent options, and a new personalized digest feature called Daily Brief. The company introduced a new AI Ultra subscription tier at $100 a month and reduced its top Gemini AI Ultra pricing from $250 a month to $200.
- •Google embedded AI agents called "intelligent search box" directly into Google Search, which now generates contextual answers, images, and short video clips in response to questions; Generative UI—custom layouts created on-the-fly based on search result type—will roll out to everyone this summer. Search agents are also being integrated into Maps (Ask Maps, released in March) and YouTube (Ask YouTube feature, which finds relevant videos and snaps to points answering specific questions).
- •Google announced Gemini Omni, an AI video generator that can modify selfie videos to add different backgrounds, styles, or environments, incorporate real video and images from a camera roll, and create cinematic effects. OmniFlash, a starter version, is available starting today for Google AI+ Pro and Ultra subscribers; full Omni still-image and text capabilities will come later.
- •Gemini Spark, a personal AI agent running on Google Cloud, can write emails, plan events, manage tasks, and pull information from Google Drive files while learning a user's schedule and life rhythms. It currently works only with Google software but will expand to Chrome and third-party support later this summer.
Related Articles
Large Language Models
Simon Willison publishes monthly briefing on LLM developments
Simon Willison's Weblog·May 20, 2026
Large Language Models
Google releases Gemini 3.5 Flash at 3× the price of its predecessor, deploying it across consumer products and APIs
Simon Willison's Weblog·May 20, 2026
Large Language Models
Google launches Gemini 3.5 Flash with 1M-token context and Gemini Omni video generation at I/O 2026; processes 3.2 quadrillion tokens/month across 900M+ monthly users
Latent Space·May 20, 2026
Large Language Models
Google restructures AI subscriptions at I/O 2026 with three tiers and consumption-based billing replacing daily prompt limits
THE DECODER·May 19, 2026
Large Language Models
Google is expanding Gemini's access to personal data across its services, from Gmail and Photos to local computer files, positioning AI features as central to its productivity strategy.
The Verge AI·May 19, 2026
Stay ahead with AI news
Get curated AI news from 200+ sources delivered daily to your inbox. Free to use.
Get Started Free