Google Gemini auto browse leak: Is this the AI browser agent that will change Chrome?

Google is reportedly developing Gemini Auto Browse, an AI browser agent in Chrome that allows Google Gemini to actively browse the web and perform real actions for users. Google is also said to be in the process of developing a new capability of Gemini which is powerful that can transform the way in which people can interact with the web. To provide active use of the Chrome browser on behalf of a user, the company is testing an AI browser agent, internally named Auto Browse, which plans to use the Chrome browser to enable Gemini to actively use the browser, according to recent leaks and code references. When released, this would take Gemini to the next level of being a conversational assistant, and become an AI capable of performing real actions on the internet.

What is Auto Browse?

Auto Browse turns Google Gemini into an AI that browses the web, enabling it to open tabs, click links, navigate websites and complete tasks directly inside Chrome. Auto Browse as a feature of Gemini can be seen by looking at early leaks that have been noticed in Chrome Canary builds and talked about by insiders on Reddit: it can be described as a tool that allows controlling a browser cursor. This implies that Gemini would be able to open tabs, follow links, browse websites and use web pages just like a human user.

This system would enable Gemini to go to sites, glean information on multiple pages and close the gap on tasks as opposed to the current AI assistants that operate based on search returns and summaries. Although Google has not publicly verified Auto Browse, its existence includes relevant code and user interface features, which is heavily suggestive of internal testing and not a theoretical test.

Why does Chrome give Google a structural advantage?

Unlike competitors, Google Chrome AI integration allows Gemini to work natively inside the browser without extensions, giving Google a major advantage in AI automation. Brower-based AI agents are also being investigated in other companies. OpenAI is experimenting with web-capable agents using experimental tools and Perplexity has shown AI-assisted web browsing experiences. The situation of Google is, however, entirely different as it is the owner of Chrome.

Since Chrome commands more than 60%  of the worldwide browser market, Google could incorporate Gemini into the browser itself and not use extensions or workarounds. This enables more system accessibility, high performance, and user-friendliness. It also helps Gemini to operate with the workflow, tabs, bookmarks and preferences.

To businesses and specialists this close integration may save a lot of time that may otherwise be wasted in the process of manual research, monitoring and repetitive Internet browsing.

Gemini ultra leaks

Leaks suggest Gemini Auto Browse could be exclusive to Gemini Ultra features, making it one of the most valuable upgrades Google has offered so far. There are rumours that Auto Browse can be limited to Gemini Ultra, the high-end AI offering of Google. In case this happens to be so, then it would be a significant change in terms of paid AI subscriptions value proposition.

Users would not be spending more money on getting more desirable answers or faster replies, but they would be spending the money on automation-AI that does the job, as opposed to merely describing it. To businesses, the researchers, marketers and analysts this type of automation at the browser level may render the upgrade worthwhile in the short term since it will save them the hours of manual work. If browser-level automation is locked behind a Gemini Ultra subscription, businesses and power users may find the upgrade pays for itself quickly.

Chrome AI automation: Practical use cases for businesses and professionals

With Chrome AI automation, Gemini could automate research, competitor tracking and data collection tasks that currently take hours of manual browsing. A Chrome-based AI browser agent can also be used to automate the process of research by browsing various sources and reading lengthy articles and enhancing this information into structured summaries. It would be able to compare products or services on different websites, monitor changes in prices or messages and generate reports automatically.

In the case of content teams, Gemini can search the web items and gather trends, statistics using trusted sources and make research notes. In the case of operations and administration, it may be useful to collect the data of various portals and pre-complete the forms, leaving the ultimate approval to the user.

This is in line with the larger vision of agentic AI that Google holds systems that plan, act, and execute tasks as opposed to reacting to single-use prompts.

Security and prompt injection risks

Because Auto Browse functions as an AI browser agent in Chrome, Google is reportedly building safeguards against prompt injection and malicious web instructions. The browser agents present new security solutions, specifically prompt injection attacks, in which rogue websites are trying to influence the behaviour of AI via covert commands. Notably, Google already published a study on how it can protect itself against prompt injection by browser-based AI agents, which suggests the risks are not unfamiliar to Google.

The integrated security model of the Chrome along with action review controls and sandboxing, puts Google in a better position to deal with these risks than most of its competitors. Nevertheless, its widespread acceptance will be determined by the level of openness with which Google will enable users to face and agree with AI-based activities.