Link: OpenAI’s new Operator AI agent can do things on the web for you

OpenAI has unveiled a research preview of an AI agent named Operator designed to perform web-based tasks. It uses an embedded browser to interact with websites by typing, clicking, and scrolling.

Initially available in the US to subscribers of the $200 monthly ChatGPT Pro tier, Operator integrates GPT-4o’s vision with advanced reinforcement learning. This allows it to interact with graphical user interfaces.

Operator can autonomously correct its actions and will transfer control to the user for complex decisions or sensitive information requests. It has safeguards to refuse harmful requests and block inappropriate content.

OpenAI is collaborating with companies like DoorDash and Uber to ensure Operator meets practical needs and adheres to social norms. However, it may struggle with complex tasks like managing calendars or creating slideshows.

Plans are underway to expand Operator’s availability to Plus, Team, and Enterprise users, integrating these features further into ChatGPT. This progression aims to enhance Operator's functionality and accessibility.

 #

--

Yoooo, this is a quick note on a link that made me go, WTF? Find all past links here.