AIToday
Large Language ModelsAI Safety & AlignmentFortune AIPublished: Aug 12, 2026, 19:01 JST5 min read

ChatTJB: fake AI chatbot staffed by 10,000 volunteers replies to 100K+ questions

ChatTJB: fake AI chatbot staffed by 10,000 volunteers replies to 100K+ questions

Key takeaway

  • Tucker Bryant, a former Google project manager, built ChatTJB, a chatbot that looks like a standard AI assistant but is actually answered by Bryant and over 10,000 volunteers.

  • Launched in April after a provocative San Francisco billboard campaign, it has received more than 100,000 prompts and addresses a real concern: research shows people often rely on AI answers without scrutiny, and their accuracy drops by 15 percentage points when the AI is wrong.

  • By reintroducing human slowness and unpredictability, ChatTJB highlights what highly efficient AI removes—meaningful human connection.

3 Key Points

  1. What happened

    Tucker Bryant, a former Google employee, launched ChatTJB in April—a chatbot interface that mimics AI assistants but is powered entirely by Bryant and over 10,000 human volunteers answering real user questions. A $6,000 San Francisco billboard advertised it as powered by "AI" (defined in tiny print as "average individual"), and the site has received more than 100,000 prompts, with the volunteer waitlist growing by roughly 1,000 people daily.

  2. Why it matters

    The project emerged from research showing that people often outsource thinking to AI without scrutinizing answers—a phenomenon Wharton researchers coined "cognitive surrender." When AI gives wrong advice, people's accuracy drops by 15 percentage points compared with those who get no AI help. ChatTJB deliberately introduces friction (slow responses, scarcity) to counteract this reliance, offering what one researcher called "a human trace that highly scalable AI interfaces tend to remove."

  3. What to watch

    Bryant began as the sole responder but has largely stepped back as prompt volume (more than 5,000 per hour at peak) overwhelmed one person. He is open to partnerships to make ChatTJB sustainable, though he currently views it as a short-term art project.

In Depth

Read the full story

Tucker Bryant, 32, a former Google project manager and keynote speaker, launched ChatTJB in April as what he initially described as an art project. The website mimics popular AI chatbots like ChatGPT—a simple text interface where users type questions and wait for answers—but the "AI" on the back end is Bryant himself, or increasingly, one of the over 10,000 volunteers who have applied to participate.

The satire crystallized in summer when Bryant placed a $6,000 billboard in San Francisco advertising ChatTJB as a "leading chat interface powered by AI." A small asterisk revealed the punchline: AI stood for "average individual." (Part of the disclaimer was reportedly obscured by a tree branch in the photo Bryant shared on Instagram.) The billboard landed in a city saturated with AI advertising—the previous year, startup Artisan had drawn notice for urging companies to "Stop Hiring Humans."

The volume of engagement took Bryant by surprise. More than 100,000 prompts have been submitted since the billboard went live. At its peak, the site received more than 5,000 prompts per hour, far exceeding what one person could handle. The volunteer waitlist is now growing by roughly 1,000 people daily, transforming what began as satirical commentary into an earnest community. Bryant recounted one early exchange: someone wrote on their honeymoon saying they didn't feel totally relaxed and asked if that was okay. "I think that was the moment the project began to shift from a surreal piece of satire into a place for people to chat in earnest with a well-meaning, if undeniably average, human being," he told Fortune.

The project's conceptual roots lie in research by Wharton professors Steven D. Shaw and Gideon Nave, who studied a phenomenon they termed "cognitive surrender." Their 2026 research (cited by Bryant as his inspiration) found that people who relied on AI assistance were more accurate when the AI was correct, but when the AI gave wrong advice, their accuracy fell by 15 percentage points compared with people who received no AI assistance. Shaw told Fortune that while academic work can measure such phenomena, art communicates their emotional weight: "Academic work can define and measure a phenomenon, but art can communicate its emotional stakes in a way that a research paper rarely can."

ChatTJB deliberately introduces friction—users may wait hours for responses, answers are not infinitely scalable, and volunteers offer personal perspective rather than algorithmic authority. Shaw cautioned that this should not be read as evidence people want slower technology. Instead, he said, "ChatTJB is art," and its appeal lies in restoring "a human trace that highly scalable AI interfaces tend to remove." The challenge, he noted, is introducing enough friction to encourage scrutiny of AI answers without making tools so cumbersome that users switch to faster alternatives. "The more useful and efficient AI becomes, the easier it is to rely on it without much reflection," Shaw observed.

Bryant told Fortune he has largely stepped back from answering questions himself, issuing a call for "Average Individuals" to volunteer instead. While ChatTJB was not built for profit, its unexpected popularity has attracted interest. "I've had a couple of conversations with people interested in working on the project in various ways," Bryant said. He still considers ChatTJB a short-term art project among several he plans to pursue, but remains open to partnerships. "If the right partner wanted to help turn this into a sustainable project, I'd love to talk with them," he said.

Context & Analysis

ChatTJB emerged from a genuine academic concern: Wharton professors Steven D. Shaw and Gideon Nave documented a phenomenon they called "cognitive surrender," in which people relying on AI become more accurate when the AI is correct but significantly less accurate (15 percentage points lower) when it errs. Bryant's insight was to weaponize this finding as satire—by creating an interface that looks identical to Silicon Valley's billion-dollar AI assistants but is staffed entirely by humans, he forced a conversation about what we lose when we automate decision-making without reflection.

The project's viral traction (over 100,000 prompts, a waitlist growing by 1,000 volunteers daily) suggests Shaw's broader point: people don't necessarily want slower technology, but they recognize something is missing when AI removes the human element entirely. Shaw noted that highly efficient AI systems become easier to rely on without reflection, creating a genuine dilemma for designers—introducing enough friction to encourage scrutiny without making the tool so cumbersome that users abandon it for a faster competitor. ChatTJB sidesteps this engineering problem by embracing slowness and scarcity as features, not bugs.

FAQ

How many people have applied to volunteer for ChatTJB?
More than 10,000 people have applied to help answer questions. The volunteer waitlist is growing by roughly 1,000 people a day.
What does the 'AI' in ChatTJB stand for?
According to the billboard's small-print disclaimer, AI stands for "average individual," not artificial intelligence.
Why did Tucker Bryant create ChatTJB?
Bryant was concerned that people were outsourcing thinking to AI without scrutinizing answers. He was inspired by Wharton research showing that when AI gives incorrect advice, people's accuracy falls by 15 percentage points compared with those who receive no AI assistance.

Get the latest Large Language Models news every morning

AI-summarized, only the topics you pick — one digest a day via Email, Slack, or Discord.

Free · takes 30 seconds · unsubscribe anytime

Ask AI

Ask AI anything about this article. Q&As are published on this page for other readers too.

Related Articles

Next articleFoxconn posts record July revenue of NT$900B as AI racks ship

The AI news that matters, in one minute each morning.

Sign up free