OpenAI may soon launch AI agents that can control your computer
OpenAI reportedly plans to release artificial intelligence (AI) agents that can perform tasks on computer systems. According to a report, the company has been working on several agent-related research projects, one of which is called the “Operator” and can perform multi-step actions on computers. The AI agents would be released in January 2025 as a research preview for developers. The company reportedly plans to access its AI agents through a native application programming interface (API) that developers can use to build software and apps.
OpenAI’s AI agents
AI agents have become a recent trend in the AI space. These are smaller AI models that have a limited but specialized knowledge base and can use specific software to perform actions such as mimicking keystrokes, clicking buttons and more. Due to the specialized nature of the models, they can perform tasks accurately and quickly.
According to a Bloomberg reportOpenAI has developed a new AI agent called Operator that can perform tasks on computers. Citing people familiar with the matter, the publication claims that users can tell the AI agent complex tasks, such as writing code or booking tickets, and it can perform them.
On Wednesday, OpenAI executives reportedly announced plans to release the tool in January 2025 as a research sample. The company would create a new developer API that would allow developers to access it.
Notably, OpenAI is reportedly working on several agent-related research projects, which are nearing completion. One such agent is said to be able to perform tasks in a web browser. Details about the other projects are currently unknown.
OpenAI CEO Sam Altman called AI agents the company’s focus during a question and answer session on Reddit earlier this month. Replying to a user, he said: “We will have better and better models. But I think agents will be the next big breakthrough.”
Anthropic, OpenAI’s competitor, released native AI agents last month. Also called computing, these agents can understand and communicate with computers, essentially allowing them to monitor and complete tasks on PCs. These agents are built on an improved version of Claude 3.5 Sonnet.
