What Google Chrome Says About the Future of Web Browsing (Premium)

When Microsoft unleashed the AI era on us all in early 2023, it caught Google by surprise, and the online giant was slow to respond. Embarrassingly, its original retort to what we now call Microsoft Copilot was a disaster and it was described internally as “rushed, botched, and myopic.” But Google quickly corrected, and like Microsoft, it initiated a quick rebranding of its initial AI offering, in Google’s case, from Bard to Gemini. Today, it is clearly a player.

Whatever one’s opinions of Gemini and Google’s general AI strategy, there are some salient facts to consider.

First, Google was an early AI innovator–among other things, it literally invented the transformer architecture that forms the “T” in GPT–and so it has a longer and deeper experience with these technologies than any of its competitors.

Google has the dominant online search service, a crucial differentiator that’s key to the success of any generative AI capabilities. It also has numerous consumer and business services with over one billion monthly active users (MAUs), giving it a Microsoft-like capability to integrate and bundle its AI into offerings that will immediately impact wide audiences.

Like Apple and Microsoft, Google makes its own web browser and uses it as a vector for distributing and promoting its other products and services, most notably search. But unlike those other companies, Google’s browser is dominant on both desktop and mobile, and this dominance means we can’t discount what it is that Google is doing in this space with AI.

Put simply, barring any antitrust invention–a remote possibility, but a possibility nonetheless–Google is ideally situated to continue or even grow its dominance across multiple products and services thanks to AI, and to establish Gemini, its in-house AI, as a dominant player in this new market. There’s a lot of talk about how AI will siphon off Google Search users, but there is zero evidence of that happening so far.

Coincidental to all this, I’ve been thinking about how AI will revolutionize the web browser–our most important app on any personal computing platform–for the past few years. Progress so far has been slow, but I think that’s tied to the dominant positions that Apple, Google, and Microsoft enjoy, as that typically triggers a defensive, conservative posture. Real innovation–and disruption–most often comes from smaller players who can move more aggressively because they have no market to protect.

And that’s what we’ve seen so far. It started with subtle changes to browser UIs, with browsers both major and minor integrating one or more third-party models into AI services, usually through a sidebar. Then we saw the first real innovation and disruption with The Browser Company, which ran into a complexity wall with its first web browser, Arc, triggering a reset that would be unthinkable for a larger, more established browser maker. Opera has since announced an AI-first browser, which we’ve still not seen, called Neon. Open AI is rumored to be making a browser, and Open AI and Perplexity have both made acquisition bids for The Browser Company, which has recovered nicely with its innovative Dia web browser. Perplexity recently announced its own web browser, Comet. And even Microsoft has pushed forward with Copilot Mode for its Edge web browser.

It’s all over the map. But it’s also all slice-in-time stuff: AI evolves quickly, so quickly, and will do so with these and other AI-powered browsers. But you have to start somewhere. I’ve looked at Perplexity Comet, Copilot Mode, and Dia so far to see if there’s any innovation happening out there–spoiler alert, there is–and I will look at other AI-powered browsers as they appear. And revisit older entries as each is improved.

But it’s finally time to look at the big dog in this fight, Google Chrome. Let’s go into this with reasonable expectations.

? The basics

Google Chrome is what most people think of–and use–when they think about a web browser. Focusing on just the desktop version for now, that means a pretty complete experience with a few notable exceptions–like a full-screen reader mode–most of which can be added with extensions. It integrates with Google accounts, of course, and supports multiple account profiles like other modern browsers. It also integrates with numerous Google services, though in my experience, any Chromium-based web browser works well in that regard. It’s just that there are certain niceties–and some tracking, of course–built-in. For example, Google Translate is available “natively” in Chrome.

Everyone gets it. This is a web browser. The web browser.

Google’s approach to adding Gemini and other AI functionality to Chrome on the desktop has been subtle to the point of uneventful. Some of the key additions include three experimental AI features–Tab Organizer, AI-based custom themes, and Help Me Write–in January 2024, bringing Help Me Write to stable in February 2024, Google Lens and natural language answers based on browsing history in August 2024, and Tab Compare and other AI-powered tabs features in September 2024. And whatever else.

But this year, Google has finally given us a peek at what it might mean to more fully integrate Gemini with Chrome.

♊ Gemini in Chrome

At Google I/O this past May, Google announced that it was adding Gemini more prominently to Chrome, via a new diamond-shaped button in the browser’s title bar, up next to the window buttons (Minimize, Maximize/Restore, and Close). This new access point is still limited to those who pay for the Google AI Pro subscription, I believe. But it will be made more broadly available, albeit only in English at first, soon.

When you click the Gemini button–or type its Alt + G keyboard shortcut–a small Gemini chat pop-up appears so you can interact with Google’s AI. This little window is movable, and it will retain its position for future launches, which is nice because it can block parts of the webpage you’re looking at. It also appears above all other windows on-screen, and not just Chrome. But otherwise, it looks and works just like the Quick assist interface we see in Microsoft Edge in Copilot Mode.

Yes, a single new button isn’t all that interesting or innovative. But there’s something going on at the platform-level here, too. By default, when you enable Gemini in Chrome in Windows, Google also adds a Gemini icon to the system tray that indicates this functionality is available even when Chrome isn’t open or running. And that means you can access Gemini in Windows at any time, using a keyboard shortcut–Alt + G in Windows, but it’s configurable–just as you can with Copilot. So as I’m typing this article in Typora, I can type Alt + G and the same little Gemini pop-up appears as you see directly in the browser.

Now, that is interesting.

To be clear, this isn’t a complete replacement for the Copilot app in Windows 11. It can’t interact with your desktop like Copilot Vision, for example. But the point is to provide more pervasive access to Gemini, even when Chrome isn’t visible. It can interact with web pages when you use it with Chrome, of course. But it’s otherwise a mini interface to Gemini, so you can still chat with the AI without the context of a particular web page.

You can also access a limited version of Gemini Live from this little window. This is similar to Copilot Voice, in that it allows you to interact with Gemini using your voice and receive spoken responses from the AI. These conversations are continuous, meaning that you don’t have to keep reactivating this feature. And on a phone or other mobile device, these conversations are multimodal, in that you can share the display or camera with Gemini Live. But through Chrome, Gemini Live is limited to the browser: You can only share the current tab with it, and there’s no way to interact with a camera or upload an image. (See below for Google’s workaround for that.)

What you can do is learn more about the web page you’re currently viewing. When I displayed Synology DiskStation DS423+ First Impressions and activated Gemini Live, a throbbing blue halo surrounded the page. And it successfully summarized the article, listed the NAS specifications, and even provided the names of a few competing products–none of which are mentioned in the article–when asked. I also received similar information when viewing a product listing on Amazon.

So that is, of course, useful. But that’s pretty much all you can do, aside from choosing a voice for Gemini Live to use and toggling live captions on/off.

? Google Lens

Google also integrates its Lens service in Chrome so you can learn more about an item in a photo you’re viewing in the browser. It’s a little tedious to use–“Search with Google Lens” is an option deep in Chrome’s main menu, but I’m not aware of a keyboard shortcut–but it did solve a problem that came up recently. When we were in Washington D.C. last week, we saw a fun mini lamp at a bar and were curious if we could buy one for ourselves. But our AI-based searches all came up blank. With Google Lens in Chrome, however, I immediately found the lamp for sale at Amazon. (And yes, I bought two.)

This type of functionality is arguably more useful, or at least more commonly needed, on mobile. But I was pretty happy to see this succeed on desktop, where my previous attempts to find this thing had failed.

?️ Other AI features in Chrome?

There are only a few other AI features in Chrome and they’re on the tame side. You can chat with Gemini directly in the Address bar by typing @gemini and then a prompt, putting it into an Ask Gemini mode. And there’s an AI-based theme generator, similar to that offered by Microsoft in Edge.

And yeah, that’s about it. But it makes sense for Google, as the dominant player in this market, to push forward slowly. It doesn’t want to confuse or upset the billions of people who use Chrome and expect it to look and work a certain way. The biggest innovative steps forward will come from smaller players first, I think.