OpenAI has introduced a new beta feature in ChatGPT, dubbed Tasks, which enables users to schedule future actions and reminders. This development marks a significant expansion of the AI chatbot's capabilities, transforming it into a more traditional digital assistant.

The Tasks feature, available to Plus, Team, and Pro subscribers, allows users to instruct ChatGPT to perform specific actions at designated times. This can include daily weather reports, reminders about important events, or even simple tasks like sending a knock-knock joke to children before bedtime. To access Tasks, subscribers need to select the "4o with scheduled tasks" option in ChatGPT's model picker and then type out their desired tasks and schedules.

The system can also proactively suggest tasks based on user conversations, although users must explicitly approve these suggestions before they are created. All tasks can be managed through chat threads or a dedicated Tasks section in the profile menu, and users will receive notifications upon task completion. However, there is a limit of 10 active tasks that can run simultaneously.

It remains unclear when or if the Tasks feature will be available to free users, suggesting that it might remain a premium feature to justify ChatGPT's subscription costs. OpenAI offers monthly $20 and $200 subscription tiers.

The introduction of Tasks represents a significant shift in ChatGPT's functionality, as it moves beyond real-time conversation and into territory traditionally held by virtual assistants. This development is part of OpenAI's broader ambitions to create more autonomous AI agents, as evidenced by the reported upcoming release of "Operator," an autonomous AI agent capable of controlling computers.

Reverse engineer Tibor Blaho has also discovered that OpenAI is working on a codenamed "Caterpillar" project, which could integrate with Tasks and enable ChatGPT to search for specific information, analyze problems, summarize data, navigate websites, and access documents – with users receiving notifications upon task completion.

The rise of "agentic" AI in 2025 is not only about technological advancement but also about economics. These agent-like features represent a strategic way to monetize expensive AI infrastructure. While OpenAI's decision to put this functionality behind ChatGPT's paywall was predictable, the real question remains: Will it deliver reliable results?

The reliability of Tasks is still a concern, as OpenAI's previous agent demos have produced inaccurate information. The coming months will reveal whether the company has addressed these fundamental reliability challenges. Additionally, questions remain about what happens if ChatGPT fails to deliver time-sensitive information.

For now, paying ChatGPT users can experiment with Tasks, but it's essential to remember that the feature is still in beta. As such, it's advisable not to rely solely on it for critical reminders just yet.