ChatGPT’s new AI agent can browse the web and create PowerPoint slideshows
OpenAI has introduced ChatGPT Agent, an AI feature that allows its assistant to complete complex tasks independently by controlling its web browser. This new capability combines functionalities from OpenAI's Operator tool and Deep Research feature, enabling ChatGPT to navigate websites, execute code, and generate documents while users oversee the process. Termed "agentic AI," this technology empowers users to delegate tasks like outfit selection, PowerPoint creation, meal planning, and spreadsheet updates to the AI agent. The system leverages web browsers, terminal access, and API connections, including ChatGPT Connectors that interface with platforms such as Gmail and GitHub.
Ars Technica•