New York City bars generative AI for its 600,000 K-8 students, the largest such ban by any US school district

New York City is banning generative AI for the roughly 600,000 students in pre-K through eighth grade -- about two-thirds of the country's largest public school system -- starting with the 2026-2027 school year. Mayor Zohran Mamdani and Schools Chancellor Emma Samuels announced the one-year moratorium September 2, framing it as a bet that human instruction and peer relationships, not chatbots, are what young children actually need to learn to think. "Children need teachers and human connection in order to learn and grow," Mamdani said. "The tech industry wants us to believe that A.I.-powered early education is not only inevitable, but necessary. We do not see it that way."

It's the largest and most restrictive policy of its kind at any U.S. school district -- not a set of guidelines or a pilot with an opt-out, but an outright prohibition on student-facing generative AI, including AI tutors and instructional software, across two-thirds of a system educating over 900,000 students in total.

Companion chatbots -- AI systems designed to simulate an ongoing relationship or offer psychological support -- are banned across every grade, including high school, where the rest of the policy is far more permissive. (The stricter line on companion chatbots specifically, holding even where tutoring tools are piloted for older students, is the one place the policy treats AI's relationship-like design as the risk itself, not just its subject matter.) That carve-out reflects a specific, named worry in the administration's own framing: unlike a tutoring tool a student opens for one assignment, a companion chatbot is built to be returned to, and city officials cited concerns about mental health and cognitive development in children specifically as the category driving that stricter line.

What's actually different by grade

The policy isn't a flat ban across the system -- it splits sharply at ninth grade, and the gap between the two halves is the real substance of what New York is testing. High schoolers keep generative AI through five closely watched pilot programs, capped at roughly 50,000 students combined, plus twice-yearly AI-literacy lessons for everyone in that age band.

Screen-time guidance layers on top of the AI rules, not instead of them: the city is recommending a 30-minute daily cap on individual devices for grades 3-5 and 45 minutes for grades 6-8, with no individual devices at all for pre-K through second grade beyond a shared classroom smart board. Those are recommendations, not mandates -- families and schools can set their own rules -- but they're the clearest signal yet of how far the administration wants to push screen exposure down for the youngest students, AI or not.

Not a ban on AI in the building

Teachers keep generative AI for lesson planning, administrative work, and family communication, provided the tools clear the education department's safety standards -- the ban targets what students touch directly, not what runs the school. Assistive technology for students with disabilities, tools for multilingual learners, and career-readiness and computer-science programs are explicitly exempted at every grade, which is the same carve-out structure most state-level AI-in-education rules have settled on this year. The city is also running a wider review of technology purchased for schools generally, beyond generative AI specifically -- Chalkbeat reported the review already touches products like electronic gradebooks and family-communication platforms, suggesting the moratorium is one piece of a broader look at what software is in classrooms at all.

"Innovation does not mean more technology, and over the next year, we will lead with evidence to make sure technology serves learning -- not the other way around." -- Chancellor Emma Samuels

Not everyone reading the order is satisfied it goes far enough in the direction it claims to. United Federation of Teachers president Michael Mulgrew praised the screen-time limits but said the policy "leaves many questions unanswered," particularly how the education department will actually verify the AI safeguards built into software the city buys, rather than assuming a purchased product is safe because it cleared procurement once. His preferred fix -- regulating AI risk at the point of purchase rather than asking individual schools to evaluate compliance after the fact -- is a narrower, more technical complaint than the ban itself, but it's the one coming from the people who'll actually have to enforce the policy day to day.

Who this actually lands on

The order arrives after what the city's own account describes as real grassroots pressure, not just a top-down policy call: thousands of people signed a petition opposing generative AI in schools, and parents raised the issue at school board meetings across the system over the past year. Gov. Kathy Hochul, whose state has been pushing its own AI-in-schools rules, backed the move: "The state has been leading the way in our efforts to keep kids focused on learning and growing."

New York's move follows a summer in which several governments moved on child-facing AI from different angles and for different stated reasons -- Norway has imposed its own restrictions on AI use by younger students, and Italy requires parental consent before anyone under 14 can use most AI products, while U.S. lawmakers separately debate an age-verification bill in Congress. None of those match NYC's approach exactly: this is a district-level operational ban tied to a single school system's procurement and classroom rules, not a national consent regime or an age-verification law, and it expires -- or converts into something else -- based on what one city's own coalition finds over the coming year, not a legislature's timeline.

That makes the real news here less the ban itself than the experiment attached to it. New York is deliberately running two different policies on either side of ninth grade, inside one school system, for one year, with a coalition tasked to study what actually happens on both sides of that line. Whether the AI industry's biggest single-district loss of the year becomes a template other cities copy, or a one-year pause that quietly lapses, depends on what that coalition reports back -- and on how the five high-school pilots it's watching most closely actually go.