ChatGPT Chrome extension updated to read and act on open tabs

By Splitfeed EditorialLast updated Oct 16, 2026, 18:20 UTC

On October 15, 2026, OpenAI updated its ChatGPT Chrome extension to let the assistant read content from a user's open browser tabs. The feature activates when a user tags Chrome inside the ChatGPT web or desktop interface, at which point the assistant can summarize pages, compare products across multiple tabs, and, through a "Computer Use" feature available in Work mode, fill out forms on the user's behalf. OpenAI released a demo showing a user asking ChatGPT to compare three tabs open with sofas under consideration for purchase. The update is available across all ChatGPT plan tiers, including free accounts. The release comes after OpenAI had previously cancelled its standalone ChatGPT Atlas browser project, suggesting the company redirected that browser ambition into extending an existing tool rather than building a separate one.

The update matters because it moves ChatGPT closer to being an ambient layer over everyday browsing rather than a separate tool a user switches to deliberately. Reading across multiple tabs simultaneously and taking actions like form-filling represents a meaningful shift in what a browser-based AI assistant can do without requiring the user to copy and paste content manually. The genuine points of dispute here are around privacy and control: the extension can access whatever is open in the browser, which raises questions about what data is read, how it is handled, and whether users have enough visibility into when the assistant is actively reading their tabs versus sitting idle.

None of the three camps have published specific reactions to this story yet. The Pro-AI camp would typically welcome this as a practical productivity upgrade that makes AI more useful in daily workflows without requiring extra steps. The Anti-AI camp would be expected to raise concerns about surveillance risk, the potential for sensitive browsing data to flow through OpenAI's systems, and the normalisation of AI access to private browsing activity. The Middle Ground camp would likely call for clearer disclosure mechanisms and granular user controls before treating this kind of ambient access as a settled, safe feature.

The details worth watching include how OpenAI clarifies its data retention and processing policies for tab content, whether regulators in the EU or elsewhere flag the feature under existing privacy frameworks, and whether OpenAI expands Computer Use beyond Work mode or to additional browsers, which would signal how central this capability is to its longer-term product direction.