September 4, 2026

New York Just Told 600,000 Kids: No AI Until Ninth Grade

The largest school system in the country put a one-year pause on AI for every student through eighth grade, banned companion chatbots in every grade, and capped screens at 45 minutes a day. The mayor's reason fits in a sentence: he hasn't seen a study that says it helps, except the ones paid for by the companies selling it.

The facts:

• On Wednesday, September 2, eight days before the first day of school, New York City announced a one-year moratorium on student-facing generative AI from 2-K through eighth grade, covering nearly 600,000 students, about two-thirds of the system (amNY, The Verge)
• Companion chatbots, the kind built for friendship or emotional support, are banned in every grade; teachers may not use AI for grading, crisis management, or decisions about graduation and promotion, but may use it for lesson planning (The Verge, Gothamist, Anadolu)
• The city will switch off the AI features in nearly 40 education programs already in classrooms, including curriculum products from Amira and HMH; where the AI can't be disabled, the product goes (amNY)
• Individual screens are banned until third grade; recommended daily caps are 30 minutes for grades three through five and 45 minutes for grades six through eight (Anadolu, The Verge)
• High schoolers get AI only in specific cases, plus two 45-minute AI literacy modules a year on what AI is and isn't, its biases, and its effect on careers (Anadolu, The Verge)
• Five vetted tools, Quill, Edia, Brisk Teaching, Playlab and Intel AI-Ready Schools, run in a limited pilot for up to 50,000 high school students under trained teachers, with time limits built in (The Verge, Anadolu)
• Mayor Zohran Mamdani, 34, who says he doesn't use AI himself: "I have yet to see a study showing that AI is beneficial for students in elementary and middle school, with the exception, of course, of research sponsored by the very corporations that stand to profit" (Fortune)
• Schools Chancellor Kamar Samuels said the policy is meant to preserve the "productive struggle" of learning (amNY)
• In April, more than 250 doctors, child development experts and advocacy groups asked the city for a five-year moratorium; the policy they got is one year, and a coalition of students, parents, unions and experts will review it during the school year (Fortune, Anadolu)
• The city's schools already run a bell-to-bell ban on phones and other personal internet devices, in place since the 2025-2026 school year; at least 39 states and Washington, D.C. now restrict phones in public schools, and in March Utah and Alabama became the first states to limit classroom screens outright (The Verge, Scientific American)
• The College Board found 84% of high school students use AI to help with schoolwork, and one in five high schools allows generative AI with no formal policy on acceptable use (Scientific American)
• The education technology industry took in $48 billion in 2024 and is projected to reach $90 billion by 2030; a Utah state investigation found 61% of educational software with data agreements shared student data with third parties and 36% with advertisers (Scientific American)
• A study of nearly 40,000 Ontario students in JAMA Pediatrics found 21% use AI chatbots for emotional support, and those students were almost twice as likely to show clinically significant distress; the lead author says that's a link, not a cause (CBC, New York Post)

Eight days before school starts, the country's biggest district told two-thirds of its students they won't be using AI this year. Not a guideline. A pause, with a date on it, and a mayor standing next to the chancellor saying the part out loud: the industry says this is inevitable, and we don't see it that way.

That's a different sentence than the one districts have been saying for two years. Since the chatbots arrived, the default in most schools has been a shrug with a policy attached. Use with caution. Yellow light. New York had that policy in March. It lasted six months.

what the ban actually covers

Read the details before deciding what you think, because the headline is blunter than the policy.

Students through eighth grade don't use generative AI in class. Teachers still can, for planning and paperwork, but not to grade a kid's work and not to decide who gets promoted. The chatbots built to be a friend are out in every grade, including high school. Nearly forty programs that already had AI baked in get the feature switched off, and if it can't be switched off, the program leaves the building. Kids with disabilities keep their assistive tools. Kids learning English keep their translation app. Coding classes and robotics carry on.

High school is where it gets interesting. Five tools, vetted for safety and data handling, run under teachers in a pilot that could reach fifty thousand students, and every one of them has a time limit built in. Two short classes a year teach what the thing is, what it isn't, and what it does to a career. The chancellor's phrase for the whole approach was "productive struggle," which is the plainest description of learning anybody in this fight has offered.

And underneath the AI rule is a screen rule that got less attention. No individual devices at all until third grade. Half an hour a day through fifth grade, forty-five minutes through eighth. In a system that spent a decade putting a tablet in every hand, that's the bigger reversal.

the study that doesn't exist

The mayor's argument rests on one claim: nobody has shown this helps young kids, except the people selling it. That's worth taking seriously, because it's mostly true and it's the reason the ban is a pause and not a verdict.

The research on classroom technology is a mess in both directions. A Wisconsin policy researcher told Scientific American this summer that the backlash was "a long time coming because we've had all these claims about what education technology can do that haven't always panned out." The industry pulling in forty-eight billion dollars a year has thirteen thousand districts to sell to and no one vetting the claims. Utah went looking and found that six in ten of the software packages with data agreements were sharing student data with third parties, and a third with advertisers.

The phone bans that came first are the closest test we have. The results are mixed in the honest sense: discipline problems rose in the first year of one large study and fell after, well-being went up later, test scores didn't move. A ban is a starting line, one youth advocate put it, not a solution.

Two hundred years ago the Luddites weren't against machines. They were against machines being used on terms they had no say in. That's a closer description of what New York did than "banning AI." The city isn't saying the tool is worthless. It's saying the terms were set by the vendor, and the terms are what a school is for.

the chatbot clause

The line that will matter most in five years is the one about companion chatbots. Every grade, no exceptions. This summer a study of forty thousand students in Ontario found one in five using an AI chatbot for emotional support, and those kids were twice as likely to be in real distress. The researcher was careful to say the bot didn't cause it. The kids who are hurting go looking for something that answers, and the thing that answers is always awake.

A school can't fix that. It can refuse to be the place where it starts, and it can teach a fourteen-year-old what the thing on the other end of the conversation actually is. That's what the two literacy modules are. Ninety minutes a year is not much. It's ninety minutes more than most districts offer.

going forward

New York is the largest system in the country, and what it does gets copied. Expect the one-year pause to show up in other districts' policies by winter, and expect the vendors to fight it the way they fought the phone bans: with studies of their own, and with the word inevitable.

The experts who asked for this wanted five years and got one. The coalition that reviews it this spring will have a real question in front of it, because the pilot is designed to produce an answer. Fifty thousand high schoolers, five tools, time limits, teachers in the room. If those kids learn more, the pause ends. If they don't, the pause was the policy all along.

Here's the part in a parent's hands. The forty-five-minute cap is a number you can use at home tonight, and the question the city asked its vendors is one you can ask your own district: which of the programs my kid uses has AI inside it, and can it be turned off?

The first day of school in New York is September 10. The pause runs through June. The study the mayor says doesn't exist is the one the city just started running on itself.

Sources: The Verge, amNY, Fortune, Anadolu Agency, Gothamist, Scientific American, CBC and the New York Post on the JAMA Pediatrics study.