An AI chatbot can now act on behalf of users thanks to a new agent feature that OpenAI has added to ChatGPT. The update, dubbed “agent mode,” is a significant advancement in AI assistant development since it enables ChatGPT to start performing tasks for users in real time rather than just responding to inquiries.
The new feature, which was unveiled on Thursday, enables ChatGPT to “think” and “act” through its own virtual computer. According to OpenAI, this enables users to make sophisticated requests such as “plan and order ingredients for a Japanese breakfast for four” or “check my calendar and brief me on upcoming meetings.”
Subscribers on the Pro, Plus, and Team plans can start the rollout right away. To provide a more potent assistant experience, the feature combines ChatGPT’s browsing and research tools, Operator and Deep Research.
Using a single prompt, OpenAI staff demonstrated how the agent could assist a user in getting ready for a wedding by recommending dress code-compliant outfits, identifying hotels with buffer days, and creating a comprehensive itinerary.
Google, which is incorporating comparable features into its Gemini AI assistant, is now directly competing with OpenAI. Although it has reportedly been delayed, Apple is also working on a more intelligent Siri that can use apps on users’ behalf.
OpenAI acknowledged that there are new risks associated with the upgrade, even though it also presents exciting opportunities. The business has restricted the agent’s capabilities, preventing them from performing high-risk operations like bank transfers.
Additionally, some actions, such as sending emails, need user supervision.
Privacy is still a major problem. When giving the agent access to personal information, OpenAI suggests that users exercise caution. For example, it is reasonable to grant calendar access for meeting coordination, but it is not necessary for tasks like clothing shopping.
OpenAI’s most recent release indicates that digital assistants are rapidly progressing from simple conversation to autonomous task management as tech giants compete to develop more powerful AI agents.
