Google Turns Chrome Into an AI Agent With Auto Browse

Google just made one of its boldest moves yet in the agentic AI space: Chrome can now browse the web for you. Announced on January 28, 2026, the new Auto Browse feature — powered by the company's latest Gemini 3 model — turns Chrome into an AI agent capable of handling complex, multi-step tasks across the open web.

What Is Auto Browse?

Auto Browse is a new capability within Chrome's Gemini sidebar. You type a request — say, "find me the cheapest flights to Barcelona for the first week of March" — and Chrome takes over. It opens tabs, clicks through pages, compares options, and reports back with results. Think of it as autofill's ambitious older sibling: instead of filling in a form field, it completes entire workflows.

According to Google's official blog post, testers have been using Auto Browse for scheduling appointments, filling out online forms, collecting tax documents, getting quotes from service providers, managing subscriptions, and even speeding up driving license renewals.

How It Works

When you launch Auto Browse from the Gemini sidebar, the AI takes control in its own tab, making "ghostly clicks" as it navigates sites on your behalf. You can watch it work in real time. For sensitive actions — such as making purchases or posting on social media — Auto Browse pauses and asks for your explicit confirmation before proceeding.

A key disclaimer reads: "Use Gemini carefully and take control if needed. You are responsible for Gemini's actions during tasks." Google is making clear that while the AI executes, the human remains accountable.

Availability and Pricing

Auto Browse is rolling out today in preview in the United States, exclusively for Google AI Pro and AI Ultra subscribers. It works on Windows, macOS, and Chromebook Plus. There's no word yet on when free-tier users or other countries will get access.

More Than Just Auto Browse

The Chrome update is broader than a single feature. Google also introduced:

A persistent Gemini side panel that stays visible across tabs, enabling multitasking without context switching
Nano Banana integration, allowing users to transform images directly in the browser using AI — no uploading or switching apps needed
Connected Apps support, linking Gemini in Chrome to Gmail, Calendar, YouTube, Maps, Google Shopping, and Google Flights for cross-app workflows
Personal Intelligence (coming soon), which will let Chrome remember context from past conversations to deliver more personalized, proactive assistance

The Bigger Picture

Google's move fits squarely into Silicon Valley's current vision: AI-driven browsing where the user directs and the agent executes. They're not alone. OpenAI has been developing Atlas, a browser built from the ground up around generative AI. Microsoft has been weaving Copilot deeper into Edge. The race to make the browser an AI-first tool is well underway.

Security Concerns Worth Noting

As WIRED points out, anyone experimenting with Auto Browse should consider the security implications. AI agents that navigate the open web are susceptible to prompt injection attacks — malicious websites could potentially trick the agent into performing unintended actions. Google has built in safeguards (like the confirmation step for sensitive actions), but the risk surface is inherently larger when an AI is clicking around the web autonomously.

What This Means

Auto Browse represents a meaningful step in making agentic AI practical for everyday users. It's not a research demo or a developer tool — it's shipping in one of the world's most widely used applications. Whether it proves reliable enough for daily use remains to be seen, but the direction is clear: the browser is becoming less of a window you look through and more of an assistant that does the looking for you.

Sources: