New York Bans AI in Schools Through Eighth Grade

New York City’s public schools, a system serving nearly 600,000 students, will be barred from using generative artificial intelligence in instruction for children from pre-kindergarten through eighth grade starting with the new school year. Mayor Zohran Mamdani and Schools Chancellor Karim Samuels unveiled the rules on September 2, 2026, a week before classes begin on September 10. New York media have described it as the strictest policy of its kind in the United States.

What led to the decision

The restrictions came after a fierce backlash to an earlier version of the guidance, released in March 2026. That draft sorted use cases by risk using a traffic-light system and focused mainly on what teachers were allowed to do, leaving the question of what students themselves could do almost untouched. Parent groups responded by launching a petition for a moratorium on AI in schools that drew thousands of signatures, and dozens of parents aired their frustration at school board meetings. More than half of the New York City Council signed a letter asking the mayor and the chancellor to pause classroom use of artificial intelligence. In May, Samuels acknowledged that the original guidance had missed the mark and promised a final, stricter version, a rewrite that took about two months and left schools little time to prepare.

Rules by age group

Children from pre-kindergarten through second grade will have no personal devices in class at all, leaving only smartboards in the classroom. Students in third through fifth grade will have personal device time capped at 30 minutes a day, while those in sixth through eighth grade get 45 minutes. Generative AI itself, meaning chatbots, AI-based digital tutors and AI-driven curricula, is banned outright for every student through eighth grade. High schoolers get access only in a limited form, through five pilot programs tied to literacy and career readiness.

Explaining the reasoning behind such a sharp line for younger students, Mamdani said:

“The tech industry wants us to believe that AI-driven early education is not just inevitable, but necessary. We don’t see it that way.”

In a statement carried by Tanjug, the mayor also flagged a review period before any broader rollout of the technology: “The city is imposing a moratorium on generative artificial intelligence, and we will spend the coming year studying this technology’s impact.”

What teachers can and can’t do

Teachers may still use AI to prepare lessons, translate materials and draft notes to parents. They are barred from using it to grade student work, monitor behavior, provide counseling in crisis situations, or draw up individualized education plans. The policy carves out exceptions to the general ban as well: standardized reading and math tests already administered on computers are unaffected, and separate arrangements apply to English-language learners and students with disabilities.

What the guidance does not spell out is enforcement. Schools have only a week to prepare before the September 10 start date, and officials have not said how the limits will actually be checked across a system with thousands of classrooms.

Teachers’ union: caution, not enthusiasm

United Federation of Teachers President Michael Mulgrew said parent complaints about AI were part of the pressure that led to the stricter rules. In a recent appearance on “Up Close,” he said parents are growing increasingly anxious about artificial intelligence because all they hear are frightening stories about it, and that they have every right to be protective of their children. He called the technology itself “a very dangerous thing” and said his message to schools was to steer clear of new products until they’ve been properly vetted. He also said teachers and principals, with just a week left before the school year starts, still don’t have clear guidance on how the policy will actually work.

What remains unresolved

The decision rewrites the rules for nearly 600,000 students, but it comes with no proof that capping screen time and banning chatbots will actually improve learning outcomes. The city isn’t claiming otherwise: announcing that the next year will be spent studying the technology’s effects amounts to admitting City Hall doesn’t have an answer yet, only an earlier, softer policy that its own chancellor has already conceded failed. Whether oversight is even feasible in a system this size, and whether the rules survive their first contact with everyday classroom life, will only become clear once classes start on September 10.