The OpenAI operator allows ChatGPT to use the network for you

OpenAI allows some users to try out a new ChatGPT feature that uses his artificial intelligence operate a web browser to book travel, buy groceries, search for bargains and do many other online tasks.

The new tool, called Operator, is an AI agent: it relies on an AI model trained on both text and images to interpret commands and figure out how to use a web browser to execute them. OpenAI claims it has the potential to automate many everyday tasks and work tasks.

The OpenAI operator follows competing releases of both Google and Anthropic they have demonstrated ones able to use the network. AI agents are widely seen as the next evolutionary stage for AI after chatbots and many companies have jumped on the hype train promoting them. In most cases, they are very limited in their capabilities and simply use a language model to automate things that would normally be done with plain software.

“AI is evolving from this tool that could answer your questions to one that is also able to take action in the world, performing complex, multi-step workflows,” said Peter Wellinder, vice president of product at OpenAI. “We’re going to see a big impact on people’s productivity, but also on the quality of work that people can do.”

OpenAI acknowledges that giving ChatGPT access to a web browser introduces new risks and says the operator can sometimes misbehave. It says it has introduced various new safeguards and plans to gradually expand the operator’s capabilities.

Welinder and Yash Kumar, product and engineering leads for OpenAI’s Computer Using Agent, say the plan is to learn from how people use the tool. They acknowledge that the tool can make unwanted bookings or purchases, but add that a lot of work has gone into making sure it asks before doing anything risky. “He will come back to me and ask for confirmations before taking steps that may be irreversible,” says Kumar.

OpenAI also released a new “system map” today outlining the issues that can occur with Operator. These include the potential for misunderstanding commands or deviating from what the user is asking; be abused by users; or be targeted by cybercriminals.

“It also poses an incredible amount of safety challenges,” says Kumar. “Because your attack vector area and your risk vector area increase significantly.”

The operator will initially be available as a “preview” for ChatGPT users with a Pro account, which costs a hefty $200 per month. The company says it plans to expand access while rolling out the tool slowly because it will inevitably make some mistakes along the way.

In several demos, Operator showed the potential for AI to take a more active role as a web assistant. The tool includes a remote web browser and a chat window to communicate with a user.

At WIRED’s request, the operator was asked to book Amtrak train travel from New Haven to Washington. He went to the correct website and entered the required information correctly to bring up the timetable, then asked for further instructions. If a user is logged into Amtrak’s website or a browser profile with credit card information stored, Operator will be able to go ahead and book a ticket — even though it’s designed to ask for authorization first.

Kumar asked the operator to reserve a table at the Beretta restaurant in San Francisco. The program went to the OpenTable website, found the right restaurant, and looked up availability before asking what to do next. OpenAI says it’s partnered with a number of popular sites, including OpenTable, to ensure Operator runs smoothly on them.

The new tool is based on OpenAI’s GPT-4o AI model, which can perceive a browser and a web page and converse over typed text. The tool includes additional training designed to help him understand how to complete tasks online. OpenAI will also provide its agent for PC use through its API.