OpenAI has introduced a new AI agent called “Operator” that can independently perform tasks in a web browser. This marks a significant advancement in AI-powered automation. For now, Operator is available as a preview only to ChatGPT Pro users in the US, but OpenAI plans to expand access to other subscription tiers and integrate it into ChatGPT in the future.
A research preview of Operator, an agent that can use its own browser to perform tasks for you. pic.twitter.com/wkBBDIlVqj
— OpenAI (@OpenAI) January 23, 2025
Operator is powered by the “Computer-Using Agent” (CUA) model, which combines the visualization capabilities of GPT-4o with advanced reasoning. This enables the agent to interact with graphical user interfaces such as buttons, menus, and forms, mimicking human actions like clicking, typing, and scrolling. Unlike traditional systems that rely on APIs, Operator interacts directly with websites, performing tasks such as filling out forms, booking travel, ordering products, and organizing reservations.
The rollout of Operator is being conducted cautiously to ensure safety and gather user feedback. Features like “Interception Mode” allow users to regain control during sensitive tasks, such as entering passwords or payment information. Operator also requests user confirmation before completing important actions and declines tasks that require complex or risky decisions, such as financial transactions. Privacy safeguards allow users to delete browsing data and opt out of model training.
Despite its potential, Operator faces limitations in handling complex workflows, such as calendar management or creating presentations. There are also restrictions on the number of simultaneous tasks to maintain system performance. OpenAI acknowledges these limitations as part of its iterative development process.
The release of Operator positions OpenAI in competition with similar AI agents from Anthropic and Google, who have introduced comparable tools. While Operator has been praised for its advanced capabilities, concerns about security risks, ethical implications, and potential job displacement remain central to public discussions.
