OpenAI Operator Expansion Signals Shift To Autonomous Workflows

OpenAI Operator Expansion Signals Shift To Autonomous Workflows

Today, we are launching Operator, an agent that can browse the web and complete tasks for you. It uses its own web browser to view web pages and interact by typing, clicking, and scrolling. Right now, Operator is in a research preview, so it has some limitations and will improve as we get feedback. Operator is one of our first agents and AIs that can handle tasks independently when given instructions.  

Ask the Operator to automate browser tasks such as filling out forms, placing grocery orders, and generating memes. By working with familiar websites and tools, Operator streamlines daily workflows and creates new ways for businesses to engage customers.  

We are starting with a small rollout to ensure a smooth launch. Operator is now available to Pro users in the US at operator.chatgpt.com. This research preview helps us learn and improve. As Operator develops, we plan to expand access to Plus, Team, and Enterprise users and add its features to ChatGPT. Now, let’s look at how Operator works.  

How Operator Works 

Operator runs on a new model called Computer Using Agent (CUA). CUA combines GPT-4O’s vision skills with advanced reasoning to work with graphical user interfaces, such as buttons, menus, and text fields you see on your screen.  

Operator views your screen content through screenshots and interacts by performing mouse clicks and keyboard inputs within its browser. This enables it to execute web-based tasks without needing special API integrations.  

If the operator encounters problems or makes a mistake, it can use its reasoning skills to resolve them. If it gets stuck and needs help, it gives control back to you, making sure the experience stays smooth and collaborative.  

CUA is new and has limitations, but already sets records in Web Arena and Web Voyager, two key browser benchmarks. Read more about Operator’s research in our blog post. Now, let’s see how to use Operator.  

How to Use 

To start, just tell the operator what you want it to do, and it will take care of the rest. You can take control of the remote browser at any time. The operator is also trained to ask you to take over tasks that require a login, payment info, or CAPTCHA solving.  

Customize the operator by adding instructions for its behavior across all sites or specific ones, such as setting flight preferences on booking.com. Save prompts for instant use on the homepage. Ideal for recurring tasks such as Instacart grocery restocks. Like browser tabs, initiate multiple operator sessions for parallel activities. For example, ordering a custom mug on Etsy while booking a campsite on Hipcamp.  

Ecosystem and Users 

Operator changes AI from a passive tool into an active part of the digital world. It helps users get things done faster and gives companies new ways to improve customer experiences and boost conversions. We’re working with companies like DoorDash, Instacart, OpenTable, Priceline, StubHub, Thumbtack, Uber, and others to ensure Operator meets real needs and complies with industry standards. We also see many ways operators can make certain tasks easier and more efficient, especially in the public sector. For example, we’re partnering with the City of Stockton to help people sign up for city services and programs more easily.  

We’re releasing Operator to a small group to gather feedback and improve quickly. This approach balances new features, trust, and safety, ensuring Operator delivers value to users, creators, businesses, and public organizations.  

Safety And Privacy 

Protecting users is our top priority. Operator includes three safeguard layers to prevent abuse and keep users in control. Operator is designed for user control. It prompts for input at key moments.  

  • Takeover mode: when sensitive information, such as passwords or payment details, is required, the operator prompts you to take over. During this mode, the operator does not collect or record anything you type.  
  • User confirmations: Before actions such as ordering or emailing, the Operator asks for your approval.  
  • Operator declines sensitive tasks, such as banking or job applications.  
  • On sensitive sites, the operator requires close supervision to quickly address mistakes.  

Operator provides simple controls for managing data privacy.  

  • Training opt-out: If you disable the improvement of the model for everyone in ChatGPT settings, the operator will not use your data for training models  
  • Transparent data management: The privacy section of operator settings lets you delete all browsing data and log out of all sites with one click. It also allows easy deletion of past conversations.  

Protections are in place to prevent websites from attempting to mislead the operator with hidden prompts, harmful code, or phishing attempts.   

Cautious navigation: operator recognizes and ignores prompt injections.  

  • A model monitors suspicious behavior and can pause tasks if it detects something wrong.  
  • Automated systems and specialists review threats and update safeguards quickly.  

Operator is designed to refuse harmful requests and block unauthorized content. Moderation systems can warn users or revoke access if rules are repeatedly violated. Additionally, review steps have been implemented to detect and address misuse. Guidance is provided on using the operator in accordance with the usage policies.  

No system is perfect, and Operator remains in a research preview. Ongoing improvements are informed by real-world feedback and thorough testing. Visit the Operator research blog’s safety section for more information.  

Limitations 

Operator is in an early research preview. It can complete many tasks but may make errors, especially with complex user interfaces such as slideshows or calendars. User feedback will inform improvements in accuracy, reliability, and safety.  

What’s Next? 

We plan to make CUA, the model behind Operator, available in the API soon. This will let developers build agents based on CUA. We’ll share a release timeline as we get more feedback from the research preview.  

Enhanced capabilities: We’ll keep working to help Operator handle longer, more complex workflows. We’ll expand Operator to support plus team and enterprise users and integrate its capabilities directly into ChatGPT in the future, once we are confident in its safety and usability at scale, unlocking seamless, real-time, asynchronous task execution.  

Source: Introducing Operator 

OpenAI introduces Operator, an AI agent built for autonomous web tasks.

OpenAI introduces Operator, an AI agent built for autonomous web tasks.

Today, we are launching Operator, an agent that can browse the web and complete tasks for you. It uses its own browser to view web pages and interact by typing, clicking, and scrolling. Right now, Operator is in a research preview, so it has some limitations and will improve as we get feedback. An operator is one of our first agents who can handle tasks independently when given instructions.  

You can ask Operator to handle repetitive browser tasks, such as filling out forms, ordering groceries, or creating memes, using the same websites and tools people already use. Operator saves time and opens new ways for businesses to connect with customers.  

Next, let’s talk about access and rollout. We’re starting with a small rollout for pro users in the US at operator.chatgpt.com. This research period lets us gather feedback and improve Operator over time. We plan to expand access to the Plus team and enterprise users and integrate these features into ChatGPT.  

How Operator Works 

The operator runs on a new model called the Computer Using Agent. CUA combines GPT-4’s vision skills with advanced reasoning using reinforcement learning. It’s trained to work with graphical user interfaces, such as buttons, menus, and text fields you see on your screen.  

The operator takes screenshots of your screen and interacts with websites using mouse and keyboard actions, so it can perform web tasks without requiring special API connections.  

If an Operator encounters problems or makes a mistake, it can use its reasoning skills to fix itself. If it gets stuck and needs help, it gives control back to you, making sure the experience stays smooth and collaborative.  

CUA is still new and has some limitations, but it has already set new records in important browser benchmarks like Web Arena and Web Voyager. You can read more about the evaluations and research behind Operator in our blog post.  

How to Use 

To start, tell the operator what you want it to do. You can take control of the remote browser at any time. The operator will ask you to take over tasks that require a login, payment info, or CAPTCHA.  

Personalize Operator with your own instructions for all or specific sites, such as airline preferences on booking.com. You can set quick-access prompts for frequent tasks and manage multiple tasks at once by starting new conversations.  

Ecosystem and Users 

Operator changes AI from a passive tool to an active helper in the digital world. It makes tasks easier for users and helps companies offer better customer experiences and improve conversion rates. When working with companies like DoorDash, Instacart, OpenTable, Priceline, StubHub, Thumbtack, Uber, and others, ensure the Operator meets real needs and complies with industry standards. We also see many ways operators can make certain workflows more efficient and accessible, especially in the public sector. For example, we are partnering with the City of Stockton to help people enroll in city services and programs more easily.  

As we continue to evaluate Operator during its research period, we aim to identify and expand on ways AI can simplify civic engagement for residents. —Jamil Niazi, Director of Information Technology, City of Stockton 

Source: Introducing Operator 

OpenAI Expands “Operator” Tools – Chat GPT Moves Beyond Chat

OpenAI Expands “Operator” Tools – Chat GPT Moves Beyond Chat

OpenAI has introduced Operator, an autonomous AI agent that directly interacts with the web, performing tasks and managing digital workflows. Now available as a research preview for ChatGPT Pro users in the US, Operator can browse the internet, click buttons, complete forms, and make purchases.  

Key Features of Operator 

The operator uses the new computer using agent (CUA) model, which combines GPT for Oz’s ability to process visual information with advanced reasoning taught through practice and feedback.  

  • The operator interacts with any site by viewing pages as screenshots and performing basic actions, eliminating the need for additional API.  
  • An operator automates tasks such as booking travel or ordering groceries online.  
  • For secure actions, the operator lets users handle sensitive steps for safety.  
  • Users can run several tasks in separate conversation threads.  
  • If an issue arises, the operator diagnoses and resolves it to complete the workflow.  

Expansion Beyond Chat 

Operator, together with features such as deep research and model context protocol (MCP), supports AgentTech AI by enabling Skywind’s solution to execute end-to-end actions. The model context protocol (MCP) is a system that extends the AI agent’s ability to manage and use contextual information during complex tasks. Deep research refers to an AI’s ability to automatically gather and summarize relevant information from multiple online sources.  

  • Workflow automation: Operator manages long-running, multi-step tasks, including gathering information, synthesizing it, and executing final actions.  

Access And Availability 

  • Current Status: Operator is available as a research preview for ChatGPT Pro users in the US at operator.chatgpt.com  
  • Future Rollout: OpenAI plans to expand these capabilities to the Plus team and enterprise users.  
  • Limitations: although a capable operator is in early development and may meet challenges on complex interfaces or non-standard websites  

An operator represents an important step toward developing AI agents that serve as personal digital assistants, saving time by automating routine web tasks.  

OpenAI CEO Sam Altman predicts 2025 will be pivotal for AI agents, which automate tasks for users.  

OpenAI is now taking its first substantial step in this area.  

On Thursday, OpenAI launched a research preview of Operator, a general-purpose AI agent that controls a web browser and performs actions independently. US users on ChatGPT’s $200 Pro plan will first access Operator, with future expansion to Plus, Team, and Enterprise users.  

“[Operator] will be in other countries soon,” OpenAI CEO Sam Altman said during a livestream on Thursday. Europe will, unfortunately, take a while.  

The initial research preview is live at operator.chatgpt.com. OpenAI plans to integrate the operator into all ChatGPT clients soon.  

According to OpenAI, operators can automate tasks such as booking, travel, making restaurant reservations, and shopping online. Users can select from categories such as shopping, delivery, dining, and travel, each with its own set of automation options. When users activate Operator, a window appears that displays a dedicated web browser the agent uses to complete tasks and provide explanations of its actions. Users maintain control of their screen while the Operator works in their own browser.  

OpenAI states that the operator uses a computer using an agent model or CUA, which combines the vision capabilities of the GPT-4.0 model with advanced reasoning. The CUA interacts with website front ends. So it does not need developer-facing APIs to access services.  

The CUA clicks buttons, navigates menus, and fills out forms on web pages like a human user.  

OpenAI collaborates with companies such as DoorDash, eBay, Instacart, Priceline, StubHub, and Uber to ensure that operators comply with their terms of service.  

The CUA model is trained to prompt users for confirmation before finalizing tasks with external side effects, such as submitting an order or sending an email. OpenAI writes in materials provided to TechCrunch that it has already proven useful in a variety of cases, and we aim to extend that dependability across a wide range of tasks.  

OpenAI notes that the CUA is not yet reliable across all scenarios.  

OpenAI notes in a support document that operators are currently unable to reliably handle many complex or specialized tasks, such as creating detailed slide shows, operating intake, integrating intricate calendar systems, or interacting with highly customized or non-standard web interface surfaces.  

To ensure safety, OpenAI requires user supervision for certain tasks, such as banking transactions, even if the CUA and operator could perform them independently. For example, users must enter credit card information themselves. OpenAI states that the operator does not collect or take screenshots of any data.  

On particularly sensitive websites, such as email, the operator requires active user supervision. Making sure users can reach, can directly catch, and resolve any potential mistakes. The models might make. OpenAI says in its support materials.  

Although these measures limit the usefulness of the operators, they help prevent unauthorized spending. Google has adopted a similar strategy with its Project Marina AI Agent, which likewise avoids entering sensitive information like credit card numbers.  

Limitations 

Operator has several significant limitations.   

The operator is subject to daily, task-independent rate limits, although it can perform multiple tasks simultaneously. OpenAI notes that dynamic limits are in place. Overall usage limits reset daily.  

Currently, the operator restricts tasks such as sending emails or deleting calendar events for security reasons, even though the CUA is technically capable of performing these actions. OpenAI intends to expand these capabilities, but no timeline has been provided.  

An operator may become unresponsive when facing complex interfaces, password fees, or capture checks. In such situations, it prompts the user to take over according to the open air  

An Agentic Future 

OpenAI has taken a more guarded approach to developing an AI agent than competitors such as Rabbit, Google, and Anthropic, likely due to concerns about the safety risks associated with the technology.  

Allowing AI systems to take actions on the web introduces risks, such as bad actors who could automate phishing scams, DDoS attacks, or perhaps purchase tickets before others. Given ChatGPT’s widespread use, it is essential for OpenAI to implement safeguards against such exploits.  

OpenAI considers the operator sufficiently safe for release in its current form as a research preview.  

The operator employs tools that seek to limit the models’ susceptibility to malicious prompts, concealed instructions, and fishing efforts. OpenAI explains on its website that a monitoring system pauses execution if suspicious activity is detected, with automated and human review. Good pipelines. Continuously update safeguards.  

As part of its broader effort, Operator is OpenAI’s most ambitious attempt at an AI agent.  

Operator introduces features that make ChatGPT more practical similar to Siri or Alexa but demonstrates greater capabilities than previous assistants.  

AI agents are positioned at the next major step in AI with the potential to transform how people use the internet and their computers, rather than only delivering and processing information; agents can, in theory, take actions and complete tasks.  

With the release of OpenAI’s first practical agent, the feasibility of this vision will soon become evident

Source: OpenAI launches Operator, an AI agent that performs tasks autonomously 

OpenAI Signals AI Could Soon Run Apps on Your Computer

OpenAI Signals AI Could Soon Run Apps on Your Computer

OpenAI is moving from a commerce system to agentic AI using an operator and its computer with a CUA model. The operator can independently control a computer to complete multi-step tasks, enabling AI to interact with websites and apps on the user’s behalf.  

With this new paradigm in mind, consider the following outline of OpenAI’s vision for autonomous computer tasks.  

To provide better context, let’s start by focusing on the first key area:  

  • An operator is an AI agent designed to take control of a user’s web browser and eventually their computer to handle repetitive or complex tasks.  
  • The operator uses a computer running the agent CUA model, which combines GPT-4’s Visual Reasoning with Reinforcement Learning. Unlike older automation tools that require API interfaces, CUA can view the screen via screenshots and interact with graphical interfaces, as a person does.  
  • An operator can fill out forms, order groceries, do research, create memes, and schedule appointments.  

The Shift to Agentic AI 

  • Operator denotes a shift from a checkbox that only talks to agents who can take action. It is built to manage long, multi-step tasks with little need for people to step in.  
  • A new ChatGPT Agent feature lets AI use a virtual computer to check calendars, book restaurants, and make slide decks.  
  • Once you set a goal, agents work on their own. For example, it could plan a weekend trip.  

Present Constraints & Safety 

  • The operator is still in the research stage and is primarily available to pro users in the US.  
  • The AI pauses for human approval before any action that can be undone, like sending emails or deleting calendar events.  

The agent can sometimes get stuck on streaky interfaces, capture or password fields, so it may need help from a person.  

Future Outlook discusses the upcoming directions and possibilities for the Operator platform and agentic AI. 

  • OpenAI plans to expand the operator to the Plus team and Enterprise users.  
  • OpenAI positions Customer Agents as a Foundation for Progress Towards Artificial General Intelligence (AGI).  
  • The aim is to move from a single tool to an ecosystem in which agents work independently across multiple systems seven times.  

The move to agentic AI is part of a broader trend in 2025, with companies like Anthropic and Google building similar capabilities.  

At the start of this year, OpenAI CEO Sam Altman predicted 2025 would be pivotal for AI agents—tools that automate tasks and act on users’ behalf.  

Building on this vision, OpenAI is now making its first real move in this area.  

OpenAI has announced a research preview of Operator, an AI agent that controls a web browser and autonomously performs tasks. It will initially be available to US users with ChatGPT’s Pro subscription and will expand to Plus, Team, and Enterprise plans, with dates to be announced.  

Operators will be available in other countries soon, though a specific launch date has not been announced. OpenAI CEO Sam Altman said during a live stream on Thursday that Europe will, unfortunately, take a while.  

Currently, the research preview is at operator.chatgpt.com. OpenAI plans to add Operator to all ChatGPT clients soon. The operator promises to automate tasks such as booking travel, making reservations, and shopping. The interface offers categories such as shopping, delivery, dining, and travel for different automations.  

When users activate Operator in ChatGPT, a dedicated web browser opens, allowing the agent to complete tasks and explain its actions. Users still control their own screen, as Operator operates in its own browser.   

OpenAI explains the browser runs on a computer using an agent or CUA model, combining GPT-4o’s vision skills with advanced reasoning. The CUA attempts to interact directly with website interfaces, bypassing developer APIs.  

This allows the CUA to click, navigate menus, and fill forms on web pages much like a person.  

OpenAI says it’s collaborating with companies like DoorDash, eBay, Instacart, Priceline, StubHub, and Uber to ensure operators comply with their terms of service.  

The CUA model is trained to ask for user confirmation before finalizing tasks with external side effects, for example, before submitting an order or sending an email, so that the user can recheck the model’s work before it becomes permanent. Open-air rights in materials provided for death crimes. It has already proven useful in a variety of cases, and we aim to extend that dependability across a wider range of tasks.  

But OpenAI warns the CUA isn’t perfect. The company says it doesn’t expect the CUA to perform reliably in all scenarios just yet.  

Currently, the operator cannot consistently handle many complex or specialized tasks. OpenAI adds support for tasks such as creating detailed slide shows, overseeing intricate calendar systems, or interacting with highly customized or non-standard web interfaces.  

To be extra careful, OpenAI requires users to supervise certain tasks, such as banking transactions, even though the CUA and operator could handle them on their own. For example, users must enter credit card information themselves. OpenAI also says the operator does not color or screenshot any data.  

On particularly sensitive websites, such as email, the operator requires active user supervision, guaranteeing users can directly catch and handle any potential mistakes the model might make, OpenAI says in its support materials.  

This does limit what an operator can do, but it also helps prevent mistakes like the agent accidentally spending your mortgage payment on edgy accent chairs. Google has taken a similar approach with its Project Marina AI agent, which also avoids entering sensitive information such as credit card numbers.  

Limitations 

The operator does have some important limitations.  

There are both daily and task-based rate limits. OpenAI says an operator can handle seven tasks at once, but there are dynamic limits on how many. There is also an AU total-usage limit that resets each day.  

For security reasons, the operator will not perform certain tasks at this stage, such as sending emails or deleting calendar events, even though the CUA can. OpenAI says this may change in the future, but there is no timeline yet.  

An operator can also get stuck if it encounters a complex interface, a password field, or a captcha. When this happens, it will prompt the user to take over.  

An Agentic Future 

Compared with competitors like Rabbit, Google, and Android, OpenAI has taken longer to develop an AI agent. This may be due to the technology’s safety risks.  

When an AI system can take actions on the web, it opens the door to much more dangerous use cases from nefarious actors. You could automate AI agents who orchestrate phishing scams or DDoS attacks or have them snatch up tickets to a concert before anyone else can. Especially for a tool as widely used as ChatGPT, it’s important that OpenAI takes steps to prepare for such exploits.  

OpenAI believes the operator is safe enough to release now, at least as a result review.  

The operator employs tools that seek to limit the model’s susceptibility to malicious prompts, consent instructions, and prompt injection. OpenAI explains on its website that a monitoring system triggers action if suspicious activity is detected, while automated and human-reviewed pipelines continuously update safety balances.  

Operator is OpenAI’s most ambitious effort so far to create an AI agent. Lastly, OpenAI launched Tasks, which gave ChatGPT basic automation features such as creating reminders and scheduling prompts to run at specific times each day. Tasks added some familiar but important features to ChatGPT, making it as practical as Siri or Alexa. However, the Operator introduces capabilities that earlier virtual assistants could not offer in AI. After ChatGPT, a new technology that will change how people use the internet and their PCs. Instead of simply delivering and processing information, agents can, in theory, take actions and actually do things.  

Now that OpenAI has released its first real AI agent, we will soon see how realistic this vision actually is.

Source: OpenAI launches Operator, an AI agent that performs tasks autonomously