ChatGPT’s new AI agent can browse the web and create PowerPoint slideshows
1 min read
Summary
OpenAI has launched ChatGPT Agent, allowing its AI assistant to complete tasks that involve multiple steps while controlling its own web browser.
The update merges capabilities from previous products, meaning ChatGPT can navigate websites, run code and create documents.
Users can ask the AI to handle requests such as creating PowerPoint slides, planning meals or updating financial spreadsheets.
The system uses a combination of web browsers, terminal access and API connections to complete these tasks and integrate with apps such as Gmail and GitHub.
The AI completes these tasks in a separate virtual computer, which has its own sandboxed virtual operating system and browser with access to the real internet, but does not control the user’s actual device.
The technology will be available for developers to use in their own products via API soon.
