OpenAI on October 15, 2026 gave its ChatGPT extension for Google Chrome the ability to read a user's open tabs and, in one of its two modes, to operate them directly - sending messages, filling out forms, downloading files and grouping tabs. The company described the update as "bringing the power of computer use to anyone who uses Chrome."

In Short

ChatGPT's add-on for Chrome can now look at the web pages someone has open and, when asked, click and type inside them to finish jobs like filling out a form. This matters for anyone who runs a website, a sign-up form or a pricing page, because an AI assistant may increasingly be the thing reading and submitting them instead of a person. The extension is free to download for every ChatGPT plan, but the sources do not say clearly whether the tab-control part works on every plan.

What OpenAI posted

The announcement came as a thread from the official ChatGPT account on X, timestamped October 15, 2026. "A few big upgrades for Chrome users today! We're bringing the power of computer use to anyone who uses Chrome," the first post reads. By the time the thread was captured, the opening post showed 1.2 million views, 866 reposts and 8,300 likes.

The thread splits the new tab features along the two modes ChatGPT now offers. According to OpenAI, "If you're in Chat, it can look across your open tabs and gather info." The more consequential capability sits in the second mode. "If you're in Work, ChatGPT can control your Chrome tabs to get things, like sending messages, filling out forms, or downloading files. It can even help you organize and group that pile of tabs you have open," according to the post.

That distinction is the core of the release. Reading tabs is retrieval: the assistant pulls text from pages already loaded in the browser and uses it to answer a question. Controlling tabs is action - the assistant clicks, types and submits inside pages where the user is typically already signed in.

A second post in the thread describes a redesigned in-browser interface. "Next up: using ChatGPT right within a Chrome tab got a major upgrade - the full power of ChatGPT is now in your browser," according to OpenAI. The post adds that "Your entire ChatGPT history is right there, and you can ask question in Chat or get things done in Work - side by side while you're on a site." A 48-second video accompanied it.

A third post offered example prompts, and a fourth linked to the Chrome Web Store listing, which describes the product in five words: "Control your browser with ChatGPT." The thread closes with a single line on eligibility: "Available for all ChatGPT plans."

What 9to5Google reported

Ben Schoon, a senior editor at 9to5Google, covered the update on October 16 at 10:50 am Pacific time. His article says OpenAI "this week announced deeper integration between ChatGPT and Google Chrome, including the ability to hand control over your tabs to the AI assistant."

According to 9to5Google, the browser add-on "now features deeper integration with Chrome, including the ability for ChatGPT to look at your open tabs and get things done with that context." The report also describes "a sidebar UI that you can use ChatGPT through, similar to how you can use Chrome's built-in Gemini integration," and embeds two demonstration videos, of 48 and 42 seconds.

Schoon's piece reproduces seven example prompts from OpenAI. They include:

  • "Compare the three couches I have open by dimensions, material, delivery time, and return policy."
  • "Compare these pricing pages against our requirements. Which plan includes everything we need?"
  • "Use the details in this event brief to fill out the registration form, and stop before submitting."
  • "Put the vendors in these tabs into my comparison sheet, with pricing, key features, and source links."
  • "Look at all the tabs I have open and compare the contractor services. Then draft an email to the one that fits my home renovation."
  • "Fill out the form I have open in @chrome ahead of seeing my new dentist. Also, can you add the appointment to my work and personal calendar."

The remaining example asks ChatGPT to explain an article "in plain English" and identify its main argument and evidence. 9to5Google confirms that the extension "is now available in the Chrome Web Store."

Where the two sources diverge

The two accounts broadly agree, but they differ on several points that matter for anyone trying to work out exactly what shipped.

Which mode controls the tabs

OpenAI's thread draws a clear line: Chat reads tabs, Work controls them. 9to5Google does not mention the modes at all. Its text quotes the tab-control sentence from the thread but presents the ability as a general feature of the extension. A reader relying on the secondary report alone could conclude that any ChatGPT user can have the assistant fill in forms, when OpenAI's own wording ties that capability to Work.

Which plans get what

"Available for all ChatGPT plans," according to OpenAI, but the post sits directly under a link to the extension download, so it reads as a statement about the extension rather than about every feature inside it. 9to5Google says nothing about plans. The ambiguity is real because of how Work was introduced. When OpenAI folded its Atlas browser into the new agent in July, ChatGPT Work began rolling out on July 9 to Pro, Enterprise and Edu subscribers, with Plus and Business following, while the desktop app reached all tiers including free. Neither the October thread nor the 9to5Google article says whether free users now get Work inside Chrome, or only the Chat side that reads tabs.

9to5Google describes a "sidebar UI." OpenAI's post refers to "using ChatGPT right within a Chrome tab" and working "side by side while you're on a site." Those descriptions are compatible with a side panel next to page content, but they are not identical, and the OpenAI post calls it an upgrade rather than a new surface. PPC Land reported in July that OpenAI was updating its Chrome extension so ChatGPT could be used inside Chrome's sidebar without installing OpenAI's own browser. The October change therefore appears to extend an existing sidebar with chat history, tab access and Work, rather than to introduce the sidebar for the first time.

Timing and examples

OpenAI's post says "today," dated October 15. 9to5Google, writing the following day, says "this week." The secondary report also lists seven example prompts, while the captured X thread shows only three before a "Show more" cut-off; the other four could not be checked against OpenAI's own post in the material available.

What neither source says

The gaps are as notable as the features. Neither the thread nor the 9to5Google report names the model behind the extension, the countries where it works, any usage limits, or what happens to the contents of tabs that ChatGPT reads. Neither mentions confirmation prompts before the assistant sends a message or submits a form. One of OpenAI's own sample prompts asks ChatGPT to fill a registration form "and stop before submitting," which suggests the stop is something the user specifies in the request; whether the product also pauses on its own is not stated.

That omission carries weight. A browser agent that can type into a signed-in email tab or a web form inherits the access of the person using it, and such access is the target of prompt injection, where hidden text on a page is read as an instruction. PPC Land documented Brave, LayerX and others disclosing prompt injection flaws in Perplexity's Comet browser between August and October 2025, with Perplexity's own security team acknowledging that malicious prompt injection remained unsolved across the industry. When Anthropic opened its own Chrome extension to a research preview of 1,000 Max subscribers in August 2025, it disclosed a 23.6% attack success rate across 123 test cases without mitigations, falling to 11.2% with them. OpenAI's October thread publishes no comparable figures.

A related risk is session hijacking, in which an attacker takes over a logged-in session rather than a password. An agent operating inside already-authenticated tabs works at exactly that layer, which is why the scope of what Work can touch, and how it is constrained, will matter to security teams before it matters to marketers.

How this fits OpenAI's browser strategy

The Chrome update is the latest step in a path OpenAI has been on for a year. The company launched Atlas, a Chromium-based browser with an agent mode limited to Plus and Pro subscribers, on October 21, 2025; at the time PPC Land noted the agent worked only in active browser tabs and needed user approval before actions ran. Less than nine months later, on July 9, 2026, OpenAI discontinued Atlas and moved its browsing functions into ChatGPT Work, an agent built on GPT-5.6, inside a unified desktop app with Chat, Work and Codex modes.

Atlas never led its category. HUMAN Security data cited in PPC Land's July coverage put Perplexity's Comet at about 47% of measured agentic web traffic in May 2026, against about 20.3% for Atlas. Moving the agent into Chrome removes the need to persuade people to switch browsers at all. Chrome held 66.282% of global web traffic in Cloudflare's Q3 2025 report, a figure PPC Land cited when covering Google's own Chrome auto browse agent for US AI Pro and Ultra subscribers in July 2026.

That puts OpenAI's agent directly inside the browser of a competitor. Google began building Gemini into Chrome in September 2025, including analysis across multiple open tabs, and in July 2026 added a persistent Gemini 3 side panel and auto browse, which Google said would pause for confirmation before purchases or social posts. 9to5Google's comparison to "Chrome's built-in Gemini integration" is apt: both assistants now offer a side panel that sees open tabs, and both can act in them. The difference is that Gemini ships with the browser, while ChatGPT arrives as an extension a user installs from the Chrome Web Store.

Why it matters for marketers and publishers

The phrase "computer use" refers to a specific technique. A computer use agent operates software the way a person does - reading the screen, then clicking, scrolling and typing - rather than calling a structured API. OpenAI's choice of that phrase signals that ChatGPT in Chrome will interact with websites through their ordinary interfaces: the same forms, buttons and checkout flows built for human visitors.

Several of OpenAI's sample prompts land squarely on commercial pages. Comparing "these pricing pages against our requirements" means an assistant, not a buyer, reading and ranking B2B pricing tables. Putting "the vendors in these tabs into my comparison sheet, with pricing, key features, and source links" turns a set of vendor sites into rows in someone else's spreadsheet. Filling a registration form from an event brief, or a dentist's intake form, means lead and booking forms may be completed by software acting for a real, signed-in person.

For analytics, that creates a measurement question with no answer yet. A form submitted by ChatGPT in Work from a user's own Chrome profile arrives with that user's cookies and browser fingerprint. Whether such traffic is distinguishable from a human session - and whether it ought to be, given that a real customer requested it - is not addressed in either source. The broader pattern of agentic AI acting on behalf of shoppers has already started to show up in bot-management data, and the Chrome extension adds a route in which the agent runs locally inside a mainstream browser rather than from a cloud data centre.

Publishers face a version of the same issue. "Explain this article in plain English. What's the main argument, and what evidence supports it?" is one of OpenAI's own examples. The page still loads in the user's tab, so the visit and any ads on it may be counted, but the reading itself shifts to a summary in the side panel. How much attention the article and its ads receive when an assistant is summarising alongside is something neither OpenAI nor 9to5Google discusses.

There is also a commercial thread running beneath the product news. ChatGPT's ad business reached a $1 billion annualized run rate by late August 2026, with about 1,200 US advertisers spending in August, according to figures PPC Land reported from Digiday and Sensor Tower. Neither source for this story says whether ads appear in the Chrome side panel. More time spent in ChatGPT while browsing other sites, though, is more time inside an environment where OpenAI now sells ads.

What remains open

Several questions follow directly from the gaps in the two sources. Which ChatGPT plans can use Work inside Chrome, given that Work launched for paid tiers in July? Does the agent stop for confirmation before sending a message or submitting a form, or only when told to? Which countries are covered? What is retained from the tabs ChatGPT reads? And how will site owners, many of whom spent the past year building defences against automated traffic, treat an agent that arrives inside a customer's own browser session?

OpenAI's thread answers none of these. The Chrome Web Store listing and later documentation may.

Timeline

Summary

Who: OpenAI, through the official ChatGPT account on X, with secondary reporting by Ben Schoon of 9to5Google. The update affects ChatGPT users on Google Chrome and the websites they visit, including advertisers, publishers and any business that runs web forms.

What: The ChatGPT extension for Chrome can now read a user's open tabs in Chat mode and, in Work mode, control them to send messages, fill out forms, download files and group tabs. The in-browser interface was upgraded to include full chat history alongside the page. OpenAI says the extension is available for all ChatGPT plans, though which plans can use Work in Chrome is not stated.

When: OpenAI posted the thread on October 15, 2026. 9to5Google reported it on October 16, 2026.

Where: In Google Chrome, through an extension distributed via the Chrome Web Store. Neither source specifies countries.

Why: OpenAI describes the goal as "bringing the power of computer use to anyone who uses Chrome." The move follows the July 2026 shutdown of OpenAI's Atlas browser and places ChatGPT's agent inside the browser with the largest global share, alongside Google's own Gemini features in Chrome. For marketers, it means more forms, pricing pages and articles may be read and acted on by an assistant on behalf of a signed-in person.