Chrome brings Gemini smarts to three new countries

If you are the type of person who leaves thirty tabs open just to compare flight prices or buy a couch, Google is trying to clean up your mess.

Google just expanded Gemini in Chrome to users in India, New Zealand, and Canada. The rollout brings the AI assistant directly to desktop users on Mac, Windows, and Chromebook Plus, as well as iOS, while adding support for over 50 new languages like Hindi, French, Spanish, Bengali, and Tamil.

An AI assistant that lives in your tabs

Built on Gemini 3.1, the feature lives in a browser sidebar that you trigger with a single click. Instead of copying text back and forth between tabs and chat windows, you can ask Gemini to summarize long chemistry papers, create study quizzes, or recall pages you browsed days ago.

The real highlight is cross-tab research. Gemini can pull context across multiple active tabs simultaneously. If you are shopping for vegan protein powder or organizing an offsite ice breaker, it can consolidate the scattered info into a clean comparison table in seconds.

Google is also weaving in its broader ecosystem. From the sidebar, you can pull information from Drive and Keep, grab location details from Maps, pull timestamped summaries from YouTube videos, and even draft and send emails via Gmail without leaving your active tab. For visual tasks, Google packed in Nano Banana 2, letting you transform images on the fly—like previewing how a piece of furniture looks in your room.

To keep things from going off the rails, Google baked in defenses against prompt injections and requires confirmation before taking actions like sending emails or booking calendar events. Notably, as TechCrunch pointed out, the browser-takeover agentic capabilities launched to U.S. subscribers in January were left out of this regional rollout.

Why it matters

We think this rollout shows Google finally flexing its biggest unfair advantage: distribution.

For everyday users, this removes the friction of jumping to a standalone AI tab. Being able to compare four open product pages or send an email without switching windows is a genuine daily upgrade that requires zero prompt engineering.

For the broader industry, the battleground has clearly shifted from raw model benchmarks to browser context. Arc, Microsoft Edge with Copilot, and dozens of startups want to be your browser AI, but none of them own Gmail, Calendar, Maps, and YouTube at this scale. Bringing that context directly into the sidebar creates a sticky loop that independent AI tools cannot easily replicate.

The trade-off, of course, is trust. Giving an AI model contextual awareness across your open tabs and access to send emails raises the stakes on security. Google is being cautious by requiring manual approval for outbound actions, but as browsers evolve into semi-autonomous workspaces, prompt injection and privacy will remain the ultimate tests.