New York City Schools Block ChatGPT — and Reverse Course Four Months Later

On 3 January 2023 Chalkbeat reported that New York City's public schools — the largest school system in the US — had restricted access to ChatGPT on school networks and devices, about a month after ChatGPT launched. The education department cited concerns about student learning and about the safety and accuracy of the content. On 18 May the same year, Schools Chancellor David Banks wrote that the response had been driven by "knee-jerk fear and risk" that overlooked the potential of generative AI. The swing from blocking to embracing, in four and a half months, became a small model of the questions generative AI put to schools.

January — "does not build critical-thinking"

According to Chalkbeat, education department spokesperson Jenna Lyle said: "Due to concerns about negative impacts on student learning, and concerns regarding the safety and accuracy of content, access to ChatGPT is restricted on New York City Public Schools' networks and devices." She added: "While the tool may be able to provide quick and easy answers to questions, it does not build critical-thinking and problem-solving skills, which are essential for academic and lifelong success."

It was not a total ban. Students could still use ChatGPT off school networks and devices, and individual schools could request access if they planned to study the technology behind it. Chalkbeat noted that it was unclear whether other districts were banning the program, but that the decision from the nation's largest school system could influence them. Its report described fears among educators that writing assignments could become obsolete and that the program could encourage cheating and plagiarism. According to Education Week, other large districts, including Los Angeles, also blocked ChatGPT.

Text that could not be told apart

One reason schools struggled was that there was no reliable way to tell AI-written text from human writing. On 31 January 2023 OpenAI released a classifier for indicating AI-written text. By OpenAI's own evaluation on an English "challenge set", it correctly flagged only 26 per cent of AI-written text as "likely AI-written", while mislabelling human-written text as AI-written 9 per cent of the time; OpenAI warned it "should not be used as a primary decision-making tool". On 20 July 2023 the classifier was withdrawn "due to its low rate of accuracy".

May — admitting the "knee-jerk" reaction

In his Chalkbeat op-ed of 18 May 2023, Banks explained that ChatGPT had been placed on the city's list of restricted websites. "This doesn't prohibit its use entirely, but it does require schools to request access for staff and students," he wrote; YouTube, Netflix and Roblox shared the classification.

The response, he wrote, "allowed principals and educators to use ChatGPT while also signaling fear and risk", and "The knee-jerk fear and risk overlooked the potential of generative AI to support students and teachers, as well as the reality that our students are participating in and will work in a world where understanding generative AI is crucial." He promised to provide educators "with resources and real-life examples of successful AI implementation in schools", to "offer a toolkit of resources for educators to use as they initiate discussions and lessons about AI in their classrooms", and to build a repository and community for schools to share findings. He also encouraged schools to use "resources developed by MIT" to explore with students how AI already affects their lives.

Education Week headlined the shift an "about-face" and quoted Leigh Ann DeLyser of the nonprofit CSforAll, who said that seeing the technology used effectively in one of the city's schools gave the system's leaders confidence they could "support less technologically inclined teachers to roll this out in a way that doesn't hurt kids' learning."

Whose problem was it?

Those who backed the block feared that homework and essays would stop working as practice in students' own thinking. Banks, explaining the reversal, pointed instead to the reality that students will work in a world where understanding generative AI is crucial. Even in January, students could use ChatGPT outside the department's networks and devices; a block could only ever cover the school itself.

Later figures show how far generative AI moved into schoolwork. In Pew Research Center surveys, the share of US teens aged 13 to 17 who had used ChatGPT for schoolwork doubled from 13 per cent in 2023 to 26 per cent in 2024. In the 2024 survey, 54 per cent said using it to research new topics was acceptable, while only 18 per cent said the same of writing essays and 42 per cent said that was not acceptable. Students, too, draw lines by use.

In Japan

On 4 July 2023 Japan's Ministry of Education, Culture, Sports, Science and Technology sent boards of education provisional guidelines on the use of generative AI in primary and secondary education. According to the Japan Science and Technology Agency's Science Portal, the guidelines neither banned nor mandated generative AI, recommended starting with limited use and pilot trials at schools able to manage the risks, and called for caution with elementary pupils. Submitting AI output as one's own work and having AI answer tests were given as inappropriate uses. The ministry published a revised Version 2.0 on 26 December 2024.

In the US, Executive Order 14277 of 23 April 2025 set out to bring AI education into schools and train teachers. The decision New York City took in January 2023 became one of the starting points of a debate that moved from keeping AI out of school to deciding how to teach it there.