Lily Janiak at the San Francisco Chronicle interviews Tucker Bryant, a “32-year-old San Francisco poet, conceptual artist and keynote speaker,” who launched ChatTJB.org, a parody chatbot styled after ChatGPT, but all queries are answered by him, an “Average Individual” (paywalled, but accessible via Reader mode):

The project, he hoped, would comment on the concept of cognitive surrender, whereby AI dependency leads users to distrust their own intuition and relinquish their critical thinking. Instead of a chatbot trawling the entire internet’s wisdom for a tidy, chipper bullet-point list in response to a query or bit of small talk, he’d just wing it. […]

But instead of merely parodying the idiom, aesthetics and media of AI companies, ChatTJB subverts them, insisting on humanity in a robot domain.

ChatTJB proved surprisingly popular—the site peaked at over 5,000 queries an hour. What Bryant describes on the site as “a short-term art project” now has him “providing crisis-response resources for when distressed users write in.”

That popularity overwhelmed its creator, leading Bryant to turn it into a “community project” with other human volunteers now responding to chats—a one-on-one improv game with random internet strangers.

That idea intrigued me enough that I signed up to be a volunteer, though Bryant is apparently inundated with such requests, so it may be a while—if ever—before I’m answering queries as a ChatTJB “AI.”

(Via Cathy Hammer.)