Chrome Can Now Use AI to Browse the Internet for You

Have you ever wanted to browse the Internet, but the idea of ​​typing a URL into your address bar seemed too exhausting? Now Google is here to help you. Today, the company announced a big expansion of its existing Gemini functionality in Chrome, with highlights including a new look for the AI ​​Companion, more integrated image editing tools, and, perhaps most impressive (but also scary), the launch of AutoNavigation, which lets Gemini take the wheel when you’re online.

The new side panel view

Credit: Google

Previously, Gemini in Chrome appeared in a small box on top of your browser, making it a bit awkward to use, especially when switching between tabs. Google’s update moves it to a drop-down side panel view that is a bit larger and won’t hide any of your other content. Instead, it will be to the right of the web page you’re viewing, so you can more easily compare the answers Gemini gives you with what you see, or continue a conversation while bouncing between multiple tabs. It will retain all the same functionality as before, including the ability to reference multiple open tabs in prompts. It’s a small change, but it should make it easier to use.

Right-click any image to edit it in Nano Banana

Credit: Google

Google’s Nano Banana image-generating AI is having a bit of a moment, and new Gemini updates in Chrome make it easier to use. Now, instead of having to download an image and re-upload it to Gemini, you can edit it using Nano Banana with a simple right-click. You can also use natural language to initiate an edit by displaying the image you want to edit on your screen and asking Gemini to edit it in the side panel. Google says it should work with just about any image you can display on the browser.

In a demo, Google showed it to journalists using a Google Photos library, but there’s nothing saying you have to stick to your own images. This immediately set off alarm bells for me, given that Elon Musk’s X is currently in hot water after allowing anyone to use Grok to edit other people’s images directly on the social media platform and without their permission. After some users started using the tool to generate explicit content from other people’s photos, it was scaled back a bit, but Google doesn’t seem concerned. When I asked about security protections for this feature, a Google spokesperson told me the following:

“We have clear policies that prohibit the use of our AI tools to generate sexually explicit content, and our tools are continually improving to reflect these policies. We invested in security from the start and added technical safeguards to help limit problematic outcomes such as violent, offensive, or sexually explicit content. »

The company hasn’t said anything about how users could use Nano Banana in Chrome to circumvent copyright, but technically the new update doesn’t really add any new features to Google’s AI image generator, it just makes it easier to access. Granted, the same also applies to the recent Grok update, and easier access can mean opening the floodgates, even if you have the best of intentions.

Use AI to automatically navigate Chrome

Credit: Google

Finally, most importantly: “Agentic” is the buzzword in AI these days, and Google doesn’t want Chrome to be left behind. So now, instead of just answering questions, Gemini can take control of your browser for you.

The feature is currently limited to Google AI Pro and Ultra subscribers, but starting today, those subscribers can ask Chrome to “auto-navigate,” performing searches, directing you to different websites, and filling out forms for you.

You can watch the AI ​​navigate the web or click another tab while it works in the background. Multiple tabs can automatically navigate at the same time, so you can multitask. The AI ​​will list the steps it takes in the side panel as it navigates, to make recording easier.

Google demonstrated this to journalists by showing the AI ​​find a specific product, go to its store page, log in to the buyer’s account (using Google Password Manager) and add it to their cart. The company also suggested that you can use automatic navigation to schedule appointments, fill out an online form using information from an uploaded PDF, collect tax documents, compare apartments listed on sites like Redfin, and more. I haven’t been able to put it into practice yet, so I can’t say how well it will perform any of these tasks, although it looked sharp in the controlled demo.

Do you really trust an AI to navigate for you?

My concern with auto-navigation is mostly with sketchy websites and permissions, although Google told me it was intended for those. Autobrowsing needs permission before it can access your Google password manager, and if it comes across a link that the AI ​​says doesn’t look quite right, it will supposedly use Chrome’s existing insecure browsing protections to navigate away. A Google spokesperson told me it’s “as secure as possible”, although I’d probably want to monitor it at least during my first few requests.

The feature also has a limitation at the moment: while it can be opened in multiple tabs at once, your auto-navigation tabs won’t be able to communicate with each other. This means that each instance of auto-navigation is isolated, but this could change in the future.

Personally, I don’t see myself using that much, especially for sensitive tasks like “collecting tax documents”, but auto-filling a basic form seems convenient. Google said automatic navigation will stop and ask you to take over sensitive steps for any tasks that might require it, like purchasing an item or submitting a form. It will not (or is not supposed to) do this last step for you, giving you the opportunity to verify how it works. In this way, it is similar to the Gemini app’s existing shopping features.

Existing and upcoming features

Gemini in Chrome can use most of Gemini’s existing features, allowing it to connect to apps like Gmail or access your chat history with the bot. But there is a major one planned for the “coming months”.

Recently, the Gemini app itself rolled out a beta for “Personal Intelligence” to paid users, allowing the AI ​​to view all your past conversations and connected apps without you having to tell it where to look. It’s essentially an extension of these existing connected apps and history features, with a reasoning model applied on top of it. For example, you could ask it to help you find new tires for your car, and it will automatically search your Gmail and Google Photos to find out what model of car you own and the last time you bought tires.

This feature is still a work in progress, but the fact that it’s even in the works means that Google is moving quickly to establish parity between all the different ways to access Gemini. All other features mentioned in this article are either already available or currently rolling out.