OpenAI unveils Operator agent to automate users tasks – what you need to know

OpenAI has unveiled Operator, its first AI agent, but the company describes it as an early research preview rather than a final product. Operator uses its own web browser to carry out tasks such as typing, clicking and scrolling, powered by a model that uses GPT-4o's vision and text reading skills. It asks for input before sensitive steps like entering passwords or payment details and before significant actions like placing an order. Currently it is available only to Pro subscribers in the US, and OpenAI is working with partners including DoorDash, Instacart and Uber to build agents with it.