Users who have subscribed to OpenAI’s ChatGPT can now request the AI assistant to set reminders or repeat tasks. Starting this week, the new task scheduling feature will be available in beta to ChatGPT Plus, Team, and Pro users globally.
Users may set simple reminders with ChatGPT, such as “Remind me when my passport expires in six months,” and the AI assistant will send a push message to whatever platform you have tasks enabled. Say something like, “Every Friday, give me a weekend plan based on my location and the weather forecast,” or “Give me a news briefing every day at 7 a.m.” and ChatGPT will respond with a recurring request.
It would appear that OpenAI’s new task function is an entry point into their AI agent model, which allows AI models to act fairly freely. The CEO of OpenAI, Sam Altman, has predicted that artificial intelligence agents will “join the workforce” in 2025, calling it a watershed year for the field. Even though Tasks is just a stripped-down agentic system, it does give users the ability to set reminders via ChatGPT, which is a functionality that most people anticipate from voice assistants like Alexa and Siri. More novel features, not seen in earlier digital assistants, are revealed by the scheduled information requests.
The “4o with scheduled tasks” option is a drop-down menu in ChatGPT that users can use to access their tasks. They can then communicate with the AI assistant through ChatGPT, specifying the desired action or reminder. According to OpenAI, ChatGPT has the ability to make task suggestions during conversations. The web app has a specialized tasks manager tab, but users may also communicate with the AI assistant through any platform to establish and manage tasks.
Now that it has a tasks feature, ChatGPT can schedule web browsing, but it won’t do background searches or purchases. If you want to watch your favorite band perform live, you can tell ChatGPT to check for tickets once a month. However, you can’t have ChatGPT purchase tickets for you or even notify you when they go on sale. However, it is a step in the direction of those systems.
OpenAI claims it is gathering user feedback during the beta phase to improve the feature before rolling it out to the public via its mobile app and free ChatGPT tier. The business has said that Advanced Voice Mode cannot be used to establish tasks for this beta launch.
Although AI assistants built on massive language models have revolutionized computing, they nevertheless fail miserably at a few basic tasks that even smartphones can handle. In order to enable their assistants to set timers and create reminders, OpenAI, Google, and other AI model developers have devised creative solutions. The stakes are low for these jobs, but OpenAI wants ChatGPT to take on far more challenging work in the future.
Reportedly, OpenAI’s upcoming release of more sophisticated agentic systems will include an agent named Operator, who can supposedly develop programming and arrange for travel. According to Bloomberg, that system might be arriving within the next week or two.
More complex agentic systems mean more room for error. Your training data goes up to October 2023. Tasks show a limited range of skills, but OpenAI’s safety measures might be challenged soon as it introduces more autonomous AI systems.