Browsers are the access point between people and the unimaginable trove of information that is the internet.
Google Chrome – by far the largest and most popular of those browsers – is on Tuesday ushering in what it’s calling “the new era of browsing” by integrating its Gemini AI across platforms like Google Search, Drive, YouTube, Calendar and Maps (to name a few) in Australia and Asia-Pacific.
Instead of opening a new tab like with other AI’s and needing to feed it all the relevant information, users will now be able to click on a simple sparkle star in the top right hand corner in order to open a side panel and chat with Google’s powerful AI assistant, Gemini 3.1.
It signals a shift in accessing information systems, for decades the internet has been governed by an unwritten rule: to get the right answer, you need to use the right search terms.
We’ve all agonised over the perfect keyword combination, or laughed at internet memes of Boomers treating Google like a crystal ball by typing out long, conversational questions (think “Dear Mr Google, is my grandson still dating that nice girl from university?”).
Ironically, those older internet users may just have been ahead of their time.
“I’ve been a Google employee for nearly 21 years,” Charmaine D’Silva, director of product management at Google Chrome and lead behind the integration told news.com.au.
“I remember when creating the right search query and the perfect prompt was an actual art form. It was really fascinating to see how everyone would try to optimise their queries to find the perfect search prompt.”
“With AI in general, I think that’s completely gone out the window.”
Ms D’Silva explained what she loves about the change is that it allows users to use their natural way of speaking and language when talking because it has context available to it.
“You can ask questions like, ‘Explain this to me,’ and the reference of what ‘this’ is completely understood by the model because it’s looking at the same page you are.”
“Now people are asking really verbose, natural questions, and that’s awesome. I think the unlocking of AI — actually having people speak in natural language — is the best unlock ever,” she said.
“It allows you to just ramble along, ask a really long query, and it will still get it right. Honestly, the longer the query, the better the responses tend to be. It’s fascinating to see that shift.”
So what does the integration actually mean?
As with the broader applications of AI the options of use for this change are vast, and it would likely be quicker to answer what it couldn’t help you with than what it could.
However as part of the advance look into the implementation, Ms D’Silva showed us three examples of how the new tool could be used.
The first was Gemini’s ability to instantly “read” and comprehend different types of media directly within the active browser tab.
Ms D’Silva, opened a technical research article, a 27-minute YouTube interview and provided a contextless image received in a text message.
Gemini explained the research article and offered comparisons to others, gave a summary and key times for the YouTube video and correctly identified the image (Justin Bieber serenading Billie Eilish to ‘One less lonely girl’ at Coachella this weekend).
The next example was demonstrating how the AI handles multi-step, logic-based chores, by running a test focused on planning a family holiday to Singapore.
Gemini combined a comparison between different travel packages on a web page with ‘personal intelligence’ (saved instructions about the user), to create a holiday that aligned with family interests and availabilities, then drafted an email to her husband with a run down of it.
For the final test, Ms D’Silva showcased the creative and shopping-focused capabilities powered by the image model ‘Nano Banana 2’ which is built into the program.
While shopping online for furniture, she found a bed frame she liked, but the surrounding room in the photo didn’t match her aesthetic. Ms D’Silva prompted Gemini to “Make this room more bohemian, but keep the bed frame the same”.
The AI analysed the image and redesigned the interior of the room in the side panel to match her personal style, retaining the exact bed frame she wanted to buy but altering the surrounds with more in keeping decorations.
Because it’s integrated with Google’s wider ecosystem, Aussies will now also be able to draft and send emails via Gmail, or schedule calendar events, without ever leaving the web page they are currently reading.
But there are limits, despite its capabilities Google is careful to emphasise the tool as an assistant, not a replacement for the web itself.
It won’t independently browse sites or complete complex actions like booking flights.
And some features, like Incognito mode support and deeper integrations with personal data, are still limited or rolling out gradually.
There are also safeguards in place, the system won’t answer certain sensitive questions, particularly around health, and includes protections against scams, malicious prompts and unsafe actions.
Ms D’Silva said the implementation was born out of looking at how people were using the web and their browser.
“We saw that people were going back and forth between multiple different tabs and multiple windows to get really, really simple tasks done.
“With that we also saw that people actually abandoned a lot of tasks because they found them to be really complicated. And so when we saw how people were using the web, we thought about how we could actually make it much more useful and simple to use as well.”
Ms D’Silva said that observation led them to the question of whether every user could actually have a personal assistant on the web to help them navigate and get things done?
“This got us to actually build out Gemini in Chrome, with the challenge of taking tasks that would take 20 different tabs and 20 different minutes, and condense them to a single tab where people can complete a task in just a few minutes.”
