Google Search is cementing its position as the ultimate testing ground and showcase for the company’s vast artificial intelligence ambitions. At the Google I/O 2026 keynote, the tech giant announced that its flagship Search engine has been upgraded to run on the newly deployed Gemini 3.5 Flash model. This architectural shift promises significantly faster inferencing, deeper contextual processing, and the native capability to interpret multiple media formats simultaneously.
Read: ENEOS creates liquid fuel from carbon dioxide and water
Marking what Google calls the most radical alteration to the search box in 25 years, the platform is introducing a brand-new Intelligent Search Box. This interface dynamically expands to accommodate complex, multi-sentence queries and accepts videos, images, document files, and even active Chrome tabs as search inputs. Rather than acting as an ephemeral repository for one-off keywords, the revamped engine allows users to engage in ongoing, multi-turn conversations to continuously refine and narrow down their web results.
While the update is a welcome milestone for power users who favour conversational AI interfaces, it represents another blow to purists who advocate for the simple accuracy of Google’s original index. Crucially, Google has dropped no hint of an opt-out toggle for these integrated capabilities, following an industry-wide trend of making AI an inescapable layer of the modern web browsing experience.
The true frontier of this update lies in the rollout of “agentic” capabilities scheduled for later this summer. Designed to execute complex workflows autonomously after an initial prompt, these advanced agents will initially target premium Gemini Pro and Ultra subscribers. Upcoming information agents, for instance, will be capable of independently scouting the web to find housing accommodations based on highly specific user constraints. According to Liz Reid, Google’s Head of Search, these digital assistants will parse everything from real-time financial, retail, and sports tickers to localized blogs and social media feeds.
Free users will also receive a curated subset of agentic updates during the winter months, notably the ability to book local hospitality experiences and services directly through the search bar. This framework includes an upgraded iteration of the duplex calling technology first previewed in 2018, allowing the AI to call local businesses to negotiate appointment bookings on behalf of the consumer.
In a move to transform Search into an all-encompassing productivity hub rather than a mere stepping stone to external websites, Google is integrating its agentic coding application, Antigravity, directly into the browser interface. Available initially to US-based Pro and Ultra subscribers, Antigravity will allow users to generate custom mini-applications on the fly. For instance, a user can instruct the search engine to build a personalized fitness tracker or compile interactive, generative UI elements to visualize complex mechanical data.
Concurrently, Google is expanding its non-subscription Personal Intelligence framework to nearly 200 countries and 98 languages. This opt-in feature allows the Gemini core to securely index a user’s Gmail, Google Photos, and Calendar data to provide hyper-individualized answers. While Google emphasizes that users maintain strict administrative control over these integrations, the aggressive rollout signals a broader corporate strategy to keep users locked permanently within the Google ecosystem.
