ChatGPT Can Now Read Open Chrome Tabs

According to a report by Stevie Bonifield of The Verge, OpenAI has made several enhancements to ChatGPT’s Google Chrome extension. With the extension installed, when Chrome is mentioned in the web version of ChatGPT or in the desktop app, ChatGPT can now reference the contents of currently open browser tabs.

OpenAI has released a demo video of this feature. The video shows ChatGPT being asked to compare three open tabs about a sofa under consideration. Work that previously required copying and pasting the contents of each tab individually is now completed simply by giving an instruction while the tabs remain open.

Features Added by the Extension

OpenAI’s published sample queries reveal the breadth of expected uses. They include operations such as using the details of an event overview to fill in a registration form and stopping just before submission. A use case is also shown for compiling vendor information from multiple tabs into a comparison table with pricing, key features, and source links.

Another example is provided in which ChatGPT reviews all open tabs, compares contractor services, and drafts an email to send to a contractor suitable for a home renovation. It also lists a use case for filling out open forms before a dental appointment and adding the appointment to work and personal calendars.

All of these are based on the idea of entrusting AI with the work of transferring information while moving back and forth among multiple tabs, through a single instruction. What stands out is that information aggregation and formatting—among the most repetitive parts of browser work—are being targeted for automation.

Persistent Sidebar and Tab Operations

In this update, ChatGPT in the sidebar has also been revamped. OpenAI explains that “all ChatGPT features” and all conversation history in Chat and Work can be used without leaving the page being viewed. The setup allows users to query AI while maintaining the screen they are working on.

Furthermore, ChatGPT Work will be able to “operate” Chrome tabs. Tasks include sending messages, filling out forms, and downloading files. It is positioned close to Google’s Gemini-based “auto browse” feature. A feature to organize and group open tabs has also been added.

The extension’s role is expanding from the stage of reading tab contents to the stage of actually performing operations on tabs. The more operations an AI can perform on behalf of a browser user, the more convenient it becomes, but it also carries the risk of erroneous operations or unintended submissions. The example shown of stopping immediately before form submission appears to be a design that takes such concerns into account.

Strategy After Atlas’s Shutdown

This move can be positioned as a step after OpenAI discontinued its own browser, Atlas. The company has shifted its focus from providing a dedicated browser of its own to entering the existing Chrome through an extension. The background is likely a judgment that offering features on a browser with many users makes it easier to secure touchpoints.

Maintaining a browser in-house entails a corresponding burden, including maintaining compatibility and security updates. With an extension, it can borrow the foundation of a supported browser while inserting only AI features. It is also a rational choice for concentrating development resources on strengthening AI capabilities.

Competing companies are exploring similar directions. Google deeply integrates Gemini into Chrome and Workspace, while Microsoft weaves Copilot into Edge and Windows. The structure in which AI assistants are involved in all kinds of work via browsers is becoming a common strategy among major companies.

Privacy and Permission Issues

The ability for AI to reference open tabs raises questions about how browsing data is handled in exchange for convenience. Which tabs, at what timing, and to what extent are read? Whether it is limited to tabs explicitly specified by the user or also includes tabs opened in the background greatly changes the effective scope of permission.

In OpenAI’s explanation, tabs are referenced when Chrome is mentioned within ChatGPT. Even so, when tab operation features are included, there remains concern that users may find it difficult to grasp the boundary between the information AI can handle and the operations it can perform. When companies introduce this for business use, they will need operational rules that define in advance the scope of information handled.

In the past, there have been reports about ChatGPT outages and the reality of human review of conversations. The more AI enters everyday browsing work, the more ensuring transparency and the appropriateness of permission design become key to adoption decisions.

Editorial Opinion

In the short term, adoption of the extension is likely to progress, especially among Chrome users. Demand is strong for delegating routine tasks such as comparing multiple tabs and transcribing information to AI, and feature competition with rivals Gemini and Copilot will intensify over the coming months.

In the long term, the trend of AI performing browser users’ work on their behalf will accelerate. Within one to three years, browsers will change in character from places to view information to places to issue instructions to AI, and the permission design of extensions will become an industry-standard issue.

From our editorial standpoint, the extent to which users can control the scope of tab reading and operation permissions will determine whether the feature spreads. We need to keep asking who is seeing what behind the convenience.

References

• “ChatGPT can peek at your Chrome tabs now”, by Stevie Bonifield — The Verge, 2026-10-15T22:43:23.000Z (ARR)
• Source URL: https://www.theverge.com/ai-artificial-intelligence/1012510/openai-chatgpt-google-chrome-tabs