Calvin Wankhede / Android Authority
TL;DR
OpenAI has announced a new ChatGPT agent mode that can operate independently without active supervision.
ChatGPT agent combines the previous capabilities of web operators and deep research, allowing it to think more like a human.
It can connect with and work on third-party apps, such as Gmail or Google Drive, with minimal input.
OpenAI has just announced that it is stepping into the era of more sophisticated agentic AI with a new ChatGPT feature. A newly launched “ChatGPT agent” will extend beyond the typical chatbot usability and perform multi-level tasks without much intervention.
The ChatGPT agent builds upon the existing operator functionality. Now, ChatGPT steers its way through multiple steps to perform tasks, but without being confined to a web browser. With the agent, OpenAI says, ChatGPT will be able to perform more complex functions while “using its own computer.” One of the key attributes of an AI agent is its ability to stop and make decisions, just as a human would. And so, OpenAI’s co-founder Sam Altman says, the agent will rely on ChatGPT’s deep research abilities, combined with its conversational skills. It will actively choose between thinking quickly or analysing something deeply based on the utility in that instant.
In one of the demo videos, OpenAI clarifies that you don’t need to be actively using the ChatGPT app or web interface to get work done, which can be beneficial while juggling multiple chores. It may need you to take over while entering your login credentials. Financial transactions are disabled out of caution, and interactions will terminate if the user moves away from the ChatGPT app to visit a payment page.
The experience is meant to be interactive, OpenAI adds, and you can interrupt it to update the instructions or add some clarification needed at any step. Using connectors, you can also access third-party apps, such as Gmail or Google Drive, and more dextrous users can link it to external apps using APIs. While boasting of the agent’s capabilities, OpenAI also lists certain limitations, noting that slideshows created from scratch may lack finesse.
ChatGPT agent is exclusively available to paid subscribers, starting with those in the “Pro” tier, whereas the Plus and Team members are slated to gain access over the next few weeks. ChatGPT’s Education and Enterprise users will receive access over the coming weeks.