The trendiest new AI chatbot is just 10,000 human volunteers manually answering your questions. 

Former Google project manager Tucker Bryant recently created an AI chatbot called ChatTJB that doesn’t use AI at all. Instead, the AI stands for “Average Individual.” 

The project started in April with him on the other side of the screen, manually answering questions. ChatTJB looked uncannily like the AI assistants Silicon Valley has poured trillions of dollars into building. A user typed a question into a plain chat box, waited a moment and got a response.

The project exploded after Bryant spent $6,000 on a San Francisco billboard promoting ChatTJB in late July. Since it went up, the site has received more than 100,000 prompts, he told Fortune, while the volunteer list has been growing by about 1,000 names a day. At the time of writing, Bryant has paused ChatTJB in response to the surge in traffic. 

The project is notable because instead of machines replacing people, people are lining up to act like machines.

One message in particular changed how Bryant viewed the project. On the first night of their honeymoon, a user wrote in to say they couldn’t fully relax and asked whether that was normal. 

Bryant said it was the moment ChatTJB began to feel less like surreal satire and more like a place where people genuinely wanted to talk to “a well-meaning, if undeniably average, human being.”

ChatTJB was never meant to be a business. But its sudden popularity has created opportunities Bryant didn’t necessarily expect.

“I’ve had a couple of conversations with people interested in working on the project in different ways,” he said. Even so, he still sees ChatTJB as a temporary art project, one of several ideas he wants to explore. 

That said, he isn’t ruling anything out. “If the right partner wanted to help turn this into a sustainable project, I’d love to talk with them,” Bryant said.

The idea for ChatTJB grew out of Bryant’s unease with how easily people can hand everyday decisions over to AI without pausing to question its answer.

He told Wired that he caught himself asking an AI chatbot if he should wear short sleeves or long sleeves on a 65-degree day in a city where he had lived for seven years. His partner had a name for that habit: “cognitive surrender.”

The term comes from research by Wharton professors Steven D. Shaw and Gideon Nave, who studied what happens when people lean too heavily on AI guidance.

“When people hear the term, they immediately recognize the phenomenon, both in others and often in themselves,” Shaw told Fortune. “Academic work can define and measure a phenomenon, but art can communicate its emotional stakes in a way that a research paper rarely can.”

The trendiest new AI chatbot is just 10,000 human volunteers manually answering your questions. 

Former Google project manager Tucker Bryant recently created an AI chatbot called ChatTJB that doesn’t use AI at all. Instead, the AI stands for “Average Individual.” 

The project started in April with him on the other side of the screen, manually answering questions. ChatTJB looked uncannily like the AI assistants Silicon Valley has poured trillions of dollars into building. A user typed a question into a plain chat box, waited a moment and got a response.