When Schools Banned ChatGPT, a Podcast Booked It as a Guest

Three months after OpenAI's ChatGPT reached the public, most American school districts wanted nothing to do with it. Students Incorporated, a weekly podcast produced by students at the International Community School of Bangkok, wanted an interview.

The resulting episode, released March 1, 2023, and titled "An Interview with AI," is a strange artifact now. Two student hosts, Lyon and Linda, email ChatGPT a formal invitation, read its reply on air, and then spend the rest of the episode questioning it about aliens, superhero capes, and the school's own list of values. It plays almost like a novelty bit and a fairly precise time capsule of the months American education couldn't decide whether generative AI was a plagiarism machine or something worth talking to, predating most of its own audience's experience with the tool. When Pew Research Center surveyed American teenagers that same year, only 13% said they had used ChatGPT for schoolwork at all; that figure would not double to 26% until the following year, Pew found in a 2025 follow-up survey.

A Guest Every Other School Was Blocking

The timing was not incidental, even if the hosts never framed it that way. ChatGPT launched on November 30, 2022, and by January it was the fastest-growing consumer application in internet history, reaching roughly 100 million users within two months, a ramp that took TikTok nine months and Instagram more than two years, according to a UBS analysis widely reported at the time. School administrators moved almost as fast, and mostly in one direction. Los Angeles Unified, the country's second-largest school district, blocked ChatGPT from its networks and devices on December 12, 2022, saying the move was meant to "protect academic honesty," according to the Washington Post. New York City's Department of Education followed in early January 2023, banning the tool across all school devices and networks; Chalkbeat, which first reported the decision, quoted a department spokesperson saying ChatGPT did not build the "critical-thinking and problem-solving skills" students needed. Seattle Public Schools banned it too, telling reporters the district "does not allow cheating and requires original thought and work from students."

ICS Bangkok's podcast did the opposite. Instead of writing ChatGPT out of the classroom, the Students Incorporated team wrote it an invitation, addressed to the chatbot by name, asking it to discuss "your past, what it's like being the most popular AI, and how you do what you do." The reply, read aloud in character on the episode, said it "would be honored to participate" and previewed a discussion of how AI was "changing the world of business and education," an unwitting preview of the exact argument playing out in American superintendents' offices that same week.

The Guest Also Filled In as Staff

The show usually includes a recurring segment where the hosts read a quote of the day and a few headline news items. For this episode, they handed that job to the guest itself. Rather than research the news, the hosts asked ChatGPT to generate four historical events tied to the episode's release date, one from each of the previous four centuries. It produced a rundown spanning North Korean missile tests in the Sea of Japan in 2016, a 1962 break between China and the Soviet Union announced by Premier Zhou Enlai, the Republic of Texas's 1836 declaration of independence from Mexico, and the 1763 Treaty of Paris ending the Seven Years' War. Whether or not every date holds up to scrutiny, letting the guest write its own segment, rather than just answer questions, is a small but telling detail: three years before "AI-generated content" became a standard newsroom disclosure, a high school podcast was already blending the two without much ceremony.

A Guest That Kept Declining to Be One

What makes the episode worth revisiting is not the novelty of booking an AI as a guest. It is how carefully that guest kept declining the premise. Asked whether it had a personality, the model demurred: it did not, "in the traditional sense of the word." Asked if it would wear a cape as a superhero, it pointed out it had no physical body and offered a virtual cape instead. Pressed on whether its abilities amounted to superpowers, it drew a sharper line than most of that era's marketing did:

"I am ultimately a tool that users can use to achieve their goals and complete their tasks."

That answer, volunteered rather than prompted, cuts against how the technology was being sold at the time, as a co-pilot, an assistant, sometimes something close to a colleague. The hosts pushed past the hedging anyway, asking the model to switch to "a female voice with an accent" for a run of jokes about guinea-pig ownership laws in Switzerland and a mathematician "afraid of negative numbers," then later asking it to compose an original poem about international-school life on the spot. It obliged both times, a small, early data point on what these systems were already reliably good at in 2023: following a task and matching a tone, all while insisting there was no self behind either.

Faith, Citizenship, and a Careful Kind of Neutrality

The strangest stretch of the interview has nothing to do with technology at all. ICS Bangkok, a Christian international school, asks graduating students to embody seven "Expected School-wide Learning Results": lifelong learner, discerning thinker, competent communicator, quality producer, team player, responsible citizen, and follower of Christ. Lyon and Linda ran the chatbot through all seven, asking for its take on each.

On lifelong learning and critical thinking, the answers could have come from any school's mission statement. On "follower of Christ," the model did something more interesting: it answered inside the frame it was handed, describing the value as "accepting Jesus Christ as one's Lord and Savior" and "seeking to live a life that is consistent with His teachings," without ever claiming a belief of its own. It was, in effect, performing fluency in one religious institution's language without taking a position, a balancing act that would become one of the defining design problems of the AI industry over the following two years, as chatbots were pushed to sound at home with wildly different audiences, a Bangkok Christian school in one conversation, a secular public university in the next, without appearing to take sides.

What Three Years Bought

The distance between this episode and the present is the real story. In 2023, according to the EdWeek Research Center's annual survey of American teachers, 34% said they used AI tools at least "a little" in their work, and only 13% had received any training on it. By 2025, teacher AI use had nearly doubled to 61%, with half of teachers reporting at least one professional-development session on the technology. New York City reversed its ChatGPT ban in May 2023, five months after imposing it, once officials concluded that blocking a tool already on every student's phone did little beyond forfeiting the conversation about how to use it well. Policy has largely moved the same direction since: more than half of U.S. states now issue official K-12 guidance built around disclosure and supervised use rather than prohibition.

None of that hindsight was available to Lyon, Linda, or the "AI language model" they had booked as a guest. What the episode captured instead was a narrow, specific window, before the debate hardened into policy, curriculum, and eventually ordinary classroom habit, when a chatbot could still plausibly be treated as a novelty booking: an invitation, a reply, a poem and a few jokes on request. Three years on, the interview reads less like a stunt and more like an early field report from a technology that had not yet decided, any more than the schools questioning it had, what kind of guest it actually was.

← Back to all posts