Former Google manager launches human-run chatbot that mimics AI, draws thousands of volunteers

Tucker Bryant, who previously managed projects at Google, unveiled ChatTJB in April as a minimalist chat interface that looks like conventional AI assistants but is operated by humans. A billboard in San Francisco advertised the platform as a leading AI chat, with a tiny note clarifying that “AI” stands for “average individual.” Since its debut, the service has logged more than 100,000 user prompts, and the volunteer roster is swelling by roughly 1,000 people each day.

At its busiest, the system handled over 5,000 questions per hour, prompting Bryant to hand over most responses to volunteers. He says the project began after he noticed his own reliance on AI for trivial decisions, a behavior scholars label “cognitive surrender.” Academic work from Wharton professors Steven D. Shaw and Gideon Nave highlights the risks of such reliance, and Bryant’s art piece aims to make users pause before accepting automated answers. While not intended as a commercial venture, the initiative has attracted interest from potential collaborators who might help sustain it.