The American research organization OpenAI has introduced a new feature for ChatGPT called Operator. This AI agent can autonomously perform actions on the internet on behalf of users.
Operator is powered by a new model called Computer-Using Agent (CUA). Combining GPT-4o's vision capabilities with advanced reasoning through reinforcement learning, CUA is trained to interact with graphical user interfaces (GUIs)—the buttons, menus, and text fields people see on a screen.
Operator can “see” (through screenshots) and “interact” (using all the actions a mouse and keyboard allow) with a browser, enabling it to take action on the web without requiring custom API integrations.
Here are some features supported by Operator:
Notably, the feature can refuse to perform certain actions that may cause harm or pose a potential threat. Operator supports almost all HTTPS protocol websites, as well as some services and applications.
When using the Operator feature, an encrypted browser is launched in the cloud. If desired, the user can take control and adjust the task. In certain situations, such as making purchases on websites, the user must confirm the request.
In the USA, owners of the premium ChatGPT Pro plan (priced at $200 per month) have already received early access to Operator. In the near future, OpenAI plans to expand support for the Plus, Team, and Enterprise plans. The feature is currently unavailable in the EU.
Previously, we reported on a new AI feature for the Reddit forum.