OpenAI Unveils 'Operator' AI for Performing Web Tasks

OpenAI has launched "Operator," an AI agent designed to perform web-based tasks autonomously using its own browser. Currently available as a research preview for Pro users in the United States, the tool aims to automate everyday activities such as filling out forms, ordering groceries, and even creating memes.

Operator represents one of OpenAI's first agents, which are AI systems capable of acting independently to accomplish specific tasks. Users can delegate assignments, such as managing online bookings or restocking household items, freeing up time for other priorities.

"Operator can interact with the web just like a human, using a browser to click, type, and scroll," OpenAI said in a statement. "It broadens the utility of AI, helping people save time on repetitive tasks while opening new engagement opportunities for businesses."

Powered by OpenAI's new Computer-Using Agent (CUA) model, Operator combines GPT-4o's advanced reasoning abilities with visual recognition capabilities to interact with graphical user interfaces (GUIs). The technology allows it to navigate buttons, menus, and text fields without requiring custom APIs.

A Research-Driven Launch

OpenAI emphasized Operator's rollout would be measured and iterative, starting small to refine the technology based on user feedback. "This research preview is crucial to learn from real-world applications and improve the system," OpenAI said. Future plans include expanding access to users on Plus, Team, and Enterprise plans and integrating Operator into the ChatGPT ecosystem.

To address privacy and user control concerns, Operator is designed to transfer tasks back to users whenever sensitive information like login credentials or payment details is needed. Users can also fully customize workflows, adding personalized instructions for specific websites and saving prompts for repeated actions.

A Vision of AI as a Digital Worker

Operator's real-world impact is supported by collaborations with major companies, including DoorDash, Instacart, OpenTable, and Uber. OpenAI is also exploring public sector applications, such as streamlining access to government services through partnerships like its pilot project with the City of Stockton.

The AI has already demonstrated record-breaking performance in WebArena and WebVoyager, two benchmarks measuring browser-use capabilities, the company said. OpenAI remains focused on fine-tuning the agent, learning from early adopters, and paving the way for wider adoption.

About the Author

John K. Waters is the editor in chief of a number of Converge360.com sites, with a focus on high-end development, AI and future tech. He's been writing about cutting-edge technologies and culture of Silicon Valley for more than two decades, and he's written more than a dozen books. He also co-scripted the documentary film Silicon Valley: A 100 Year Renaissance, which aired on PBS.  He can be reached at [email protected].

Featured

  • workshop participants discuss sustainability in open science and research

    Open Source: Advancing Our Digital Commons

    IT leaders are recognizing the benefits of a return to open strategies. CT asked Jack Suess, VP of IT and CIO at UMBC, for his views on returning to the digital commons of open source.

  • Digital cyberspace with particles and Digital data

    Report: AI Is Moving Faster than Data Trust

    AI agents are already in use or pilot at most organizations, but data visibility, governance and precision recovery capabilities have not kept pace, according to Veeam's new Data & AI Trust Gap report.

  • AI logo near computer equipment

    White House Releases National Policy Framework for AI

    The White House has released a four-page AI policy framework aimed at setting a national approach to AI, with priorities including child safety, intellectual property protections, truth and accuracy guardrails, and worker training for an AI-driven economy.

  • Binary code flows through a digital pathway with red and blue lights in a dark background

    Survey: Enterprises Say They Are Ready for Agentic AI Failures, but Few Test Recovery Often

    Most enterprise organizations say they are ready to recover from disruptions involving agentic AI, but a new survey of more than 300 IT decision-makers from Australia, New Zealand, Europe, the United Kingdom, and the United States suggests relatively few test those plans often enough to prove it.