OpenAI plans to release Operator AI agent in January – it will be able to control PCs without a user

OpenAI is preparing to release a new AI agent, codenamed Operator, which will allow it to perform a variety of tasks on the user’s computer. The release of this tool could take place as early as January 2025.

Image source: OpenAI

In the first phase, the solution mentioned will be available as a research preview via the developer API. Operator is designed to compete with similar AI agents such as Anthropic’s recently introduced Computer Use and Google’s upcoming AI agent aimed at the consumer market.

OpenAI aims to create a universal tool that can perform various operations in a web browser and is suitable for solving everyday problems of users. However, as Bloomberg notes, it is not yet known whether Operator will offer users significant advantages over similar solutions from competitors. A research version of the tool will be useful in assessing its potential benefits and identifying areas for improvement based on feedback.

The launch of Operator coincides with the publication of an OpenAI document containing recommendations for the US government on AI development strategy. This document proposes the creation of “economic zones” for the active development of AI infrastructure, as well as the formation of alliances with US allies, which will strengthen the country’s position in the technological race with China.