Google has begun placing its Gemini assistant inside the Chrome browser for people in India, Canada and New Zealand, extending a tool it has already offered in other markets. The idea is quietly ambitious: that a person need not leave the page they are reading to summarise it, compare it with others, or act on what they find. It reaches desktop users on Mac, Windows and Chromebook Plus machines, and for now it is free. Behind the convenience lies an older unease about assistants that can be quietly steered by the very pages they read, and whether the safeguards Google describes will hold once
Google Turns Chrome Into an AI Browser, Bringing Gemini to Users in India, Canada and New Zealand
Gemini must get your confirmation before sending an email or adding an event
Mark
So what actually changed here? Gemini in Chrome already existed.
Mimi
The reach changed. It's now rolling out in India, Canada and New Zealand, on Mac, Windows and Chromebook Plus. The pitch is doing things without leaving your tab.
Mark
What can it do?
Mimi
Summarise pages, make quizzes from study material, find pages you visited before, and pull from Gmail, Maps, Calendar and YouTube. It can also read several tabs at once and give you one comparison table, and generate images with Nano Banana 2.
Luke
Those are Google's own examples, though. The protein powder comparison is an illustration, not a test result. Nothing here tells us how accurate the summaries are.
Mark
And Android users?
Mimi
No direct integration in the browser yet. They can hold the power button to bring up Gemini while in Chrome.
Mark
The language support seems significant for India.
Mimi
Gemini 3.1 adds more than 50 languages, Hindi among them, which matters for a market like that.
Luke
Support is claimed, but the source doesn't say how well each language performs. Listing a language isn't the same as it working well.
Mark
What about safety?
Mimi
Prompt injection is the main worry. Google says Gemini is trained to spot known threats and asks for confirmation before sensitive actions like sending email or adding calendar events.
Luke
Note the word known. That covers threats already identified. Google also describes its red-teaming and auto-updates in its own blog, and no independent evaluation is cited. And free is described as current, not guaranteed.
O Pulso
- Gemini now sits one click away in the top-right corner of Chrome, opening a sidebar for users in India, Canada and New Zealand.
- Tabs multiply and information scatters, and the assistant is being asked to gather it into a single comparison table.
- Google is guarding against hidden instructions on web pages by training Gemini to spot prompt injections and asking for confirmation before it sends emails or adds calendar events.
- Rivals such as ChatGPT Atlas and Perplexity Comet have already faced questions about these same risks, so the rollout is being watched closely.
- Android users cannot yet reach Gemini inside the browser, and the free pricing is described as current rather than permanent.
Google has begun placing its Gemini assistant inside the Chrome browser for people in India, Canada and New Zealand, extending a tool it has already offered in other markets. The idea is quietly ambitious: that a person need not leave the page they are reading to summarise it, compare it with others, or act on what they find. It reaches desktop users on Mac, Windows and Chromebook Plus machines, and for now it is free. Behind the convenience lies an older unease about assistants that can be quietly steered by the very pages they read, and whether the safeguards Google describes will hold once real accounts and real habits meet them.
Google is bringing its Gemini assistant into Chrome for users in India, Canada and New Zealand. The company has offered the feature in other places for some time, and this expansion reaches people on Mac, Windows and Chromebook Plus devices. For now, there is no charge.
The central promise is that people will not have to leave the page they are on. Gemini appears as an icon in the browser's top-right corner and opens a sidebar where it can summarise a page, build a quiz from study material, or help someone find a tab they opened earlier, a feature aimed at those who keep many tabs open at once. It also links to Gmail, Maps, Calendar and YouTube, so a video's key points can be pulled out without watching it through.
Gemini can also draw on several open tabs and merge what it finds. A shopper might compare protein powders across retailers, or a traveller might weigh destinations, with results arriving in one table. Chrome now includes Google's Nano Banana 2 image model, so a typed prompt in the sidebar can produce a picture.
Availability varies. Desktop and iOS users get the feature first, while Android users cannot yet use Gemini inside the browser, though they can summon it from Chrome by holding the power button. The update runs on the Gemini 3.1 model, which supports more than 50 languages, including Hindi, French and Spanish.
Security is the question that follows. AI browsers have been criticised for prompt injection, where hidden instructions on a web page try to push an assistant into acting against its user. Google says Gemini has been trained to recognise such threats and must get the user's confirmation before sensitive steps like sending an email. It also says it uses automated red-teaming and that Chrome's auto-update lets it ship fixes quickly.
What matters next is how these protections perform once many people in the new markets rely on the tool with real accounts. So far the protections are Google's own account, and the free pricing may not last.
Google is putting its Gemini assistant inside Chrome for users in three more countries: India, Canada and New Zealand. The company has offered the feature elsewhere for some time. The expansion reaches people on Mac, Windows and Chromebook Plus devices, and for now it costs nothing to use.
The idea is that people no longer have to leave the page they are on. Gemini shows up as an icon in the top-right corner of the browser. Click it and the assistant opens in a sidebar, where it can summarise a web page, build a quiz from study material, or help someone find a page they visited earlier. That last function is aimed at people who keep many tabs open at once.
The assistant also connects to other Google products, including Gmail, Maps, Calendar and YouTube. Google's example is pulling the key points out of a YouTube video directly in the sidebar, without watching it through.
A further feature lets Gemini draw on several open tabs at the same time and merge what it finds. Google says a shopper could compare vegan protein powders across different retail sites, or someone planning a trip could research destinations, and the results would arrive in a single comparison table. Image generation is built in as well: Chrome now includes Google's Nano Banana 2 model, so a typed prompt in the sidebar can produce a picture.
Availability differs by platform. The feature arrives first for desktop and iOS users. Android users do not yet get Gemini inside the browser interface. They can still reach the assistant while in Chrome by holding the power button, which launches Gemini. The update runs on the Gemini 3.1 model, which adds support for more than 50 languages, among them Hindi, French and Spanish.
Security is the obvious question. AI-powered browsers have drawn scrutiny for prompt injection, in which hidden instructions on a web page try to steer an assistant into doing something its user never asked for. Several competitors, including ChatGPT Atlas and Perplexity Comet, have faced questions over such risks.
Google says Gemini in Chrome has been trained to recognise known threats, prompt injections among them. It has also added a check: the assistant must get the user's confirmation before taking sensitive actions such as sending an email or adding a calendar event. In its blog post, Google said it uses automated red-teaming to keep strengthening what it calls Chrome's layered defences, and that Chrome's auto-update lets it push fixes quickly when new threats appear.
What to watch is how those safeguards hold up once large numbers of people in new markets begin using the assistant on real accounts. The protections are Google's own description so far, and the free pricing is described as current rather than permanent.
Citações Notáveis
Google uses automated red-teaming to continually strengthen Chrome's layered defences, and auto-update lets it get fixes out quickly.— Google, in its blog post (paraphrased)
Quer a matéria completa?
Leia o original em Livemint ↗
Technology