Google’s New Chrome ‘Auto Browse’ Feature Aims to Navigate the Web on Your Behalf

Google introduced a new “Auto Browse” feature for Chrome on Wednesday. This tool, powered by Google’s latest Gemini 3 generative AI model, acts as an AI assistant designed to handle tasks in your Chrome browser, such as booking flights, searching for apartments, and managing expenses.

The launch of Auto Browse is a part of Google’s ongoing effort to integrate AI features into Chrome. Following last year’s release of the “Gemini in Chrome” mode, which provided answers based on web pages and consolidated information from multiple tabs, this is the next step in AI enhancement.

Users can access Auto Browse by using the Gemini sidebar in Chrome, but currently, it is exclusively available today in the US for subscribers of Google’s monthly AI Pro and AI Ultra plans. The timeline for broader access to nonpaying users and other countries remains uncertain.

Google’s rollout aligns with Silicon Valley’s vision for the future of web browsing, which emphasizes increased AI reliance and minimized user intervention. Options available to consumers, whether fully designed around generative AI like OpenAI’s Atlas or enhanced with AI features such as Google’s Chrome, now predominantly incorporate some form of integrated AI. (The Vivaldi browser stands out for those seeking to avoid AI-driven web experiences.)

In a prelaunch demonstration, Charmaine D’Silva, a product management director for Chrome, illustrated how Auto Browse assists with online shopping. “Instead of having to recall where I purchased an item and attempting to reorder it,” she explained, “I can now delegate the task to Auto Browse within Gemini to procure jackets for me.” She demonstrated by typing a request to the Gemini sidebar to have the bot reorder a jacket from last year and find a discount coupon code before completing the purchase.

Auto Browse in action.

Courtesy of Google

Upon activation, Auto Browse takes command of Chrome and performs automated clicks in its own tab while attempting to fulfill the assigned task. A disclaimer on the demo version warns: “Use Gemini cautiously and take control if necessary. You are accountable for Gemini’s actions during tasks.” Even as it navigates the digital landscape, Google emphasizes user responsibility for the bot’s online activities following user commands.

Currently, automation has its limitations. for sensitive tasks like posting on social media or processing payments, user oversight is still required. In such cases, the Chrome bot will present the steps it has taken and ask the user if they wish to proceed.

Those interested in trying out Auto Browse should carefully consider the security implications of this automation. Despite Google’s efforts to enhance safety, Auto Browse and similar AI tools remain vulnerable to prompt injection attacks from malicious websites, which can manipulate the bot into performing unintended actions.

This week, I’ll be exploring Auto Browse to evaluate its initial strengths, weaknesses, and what it signifies for average Chrome users. Generally, I approach agentic AI tools designed to streamline tasks with skepticism, as they often fall short of expectations and reliability. Nonetheless, Google is keen on reshaping the web browsing experience around AI. Given the company’s history of gradually rolling out new features, anticipate Auto Browse to expand its availability soon.