Two years ago, OpenAI launched the chatbot craze with the launch of ChatGPT. Now he hopes to spark interest in a new wave of artificial intelligence technology.
On Thursday, OpenAI unveiled a tool called Operator that can connect to the internet and autonomously perform tasks, such as purchasing food or making a reservation at a restaurant.
“You can browse websites and perform actions on websites, just like you and I do,” OpenAI product and engineering lead Yash Kumar said in an interview.
artificial intelligence researchers call this type of technology an ai agent. While chatbots can answer questions, write poems, and generate images, agents can use other software on the Internet.
During a briefing with The New York Times, Kumar showed how the system could make a reservation at a San Francisco restaurant through the OpenTable website and purchase a grocery list through Instacart. The operator looks and behaves very similar to ChatGPT and other chatbots. The user types a request in a small window. Then the system responds as best it can.
The user can watch as the tool opens a web browser and visits specific sites. The operator can make mistakes. But in some cases, you can fix these errors. During the demonstration for The Times, the system mistakenly assumed Mr. Kumar was in Iowa, before correctly finding a restaurant in San Francisco.
The operator is not completely autonomous. Sometimes a user needs to correct their errors and provide additional requests and suggestions. For sites like OpenTable and Instacart, users must provide their private usernames and passwords. But OpenAI said it does not store this private information.
However, the company captures data that shows how the system interacts with users and accesses sites on their behalf. You can use this data to train future versions of Operator.
OpenAI said that starting Thursday, Operator will be available to anyone who has subscribed to ChatGPT Pro, a $200-a-month service that provides access to all of the company's latest tools. It plans to offer the tool through other paid services and eventually implement it in the free version of ChatGPT. Users in the United States will be the first to receive the new tool.
(The New York Times has sued OpenAI and its partner, Microsoft, accusing them of copyright infringement of news content related to artificial intelligence systems. OpenAI and Microsoft have denied those claims.)
In recent months, other leading companies, including Google and Anthropic, have introduced similar tools. However, many of these tools are not yet widely available.
The operator is based on the same technology that underpins ChatGPT. This technology is what ai researchers call a neural network: a mathematical system that can learn skills by analyzing huge amounts of data.
Newer versions of this technology learn from a wide range of data, including text, images and sounds. In this case, Operator learned from images that show how people use spreadsheets, shopping sites and other online services. After identifying patterns in this data, the new system can use similar services on behalf of computer users.
Kumar acknowledged that, like ChatGPT and other chatbots, Operator remains an experimental technology. But he said he would continue to improve in the coming months.
“This is not the most solid thing in the world,” he said. “But it's a lot better than what this kind of technology used to be.”
