New York bars generative AI for 600,000 pupils through eighth grade

New York City will bar student-facing generative artificial intelligence in public schools from 2-K through eighth grade for the 2026-27 year. Mayor Zohran Mamdani and Chancellor Kamar H. Samuels announced the rule on 2 September. It covers about 600,000 children, two-thirds of enrolment in the largest school system in the United States.

Companion chatbots, the products sold as emotional support, are banned in every grade. Teachers may not use AI to grade student work. Individual devices are kept away from children through second grade, though classroom screens and smart boards remain. Recommended caps after that are 30 minutes a day on personal devices in grades 3-5 and 45 minutes in grades 6-8.

High school is treated differently. Students there may use AI in narrow cases: literacy support, career and certification programmes, and a pilot in a handful of classrooms. The named tools in the pilot, according to city briefings, include Quill, Edia, Brisk Teaching, Playlab and Intel AI-Ready Schools. Each is supposed to keep the student as the primary thinker and to run on a timer. All high-schoolers will sit two 45-minute modules a year on what the systems are, where they fail, how they embed bias, and what they do to jobs.

Mamdani’s sentence was the one the city wanted quoted: children need teachers and human connection in order to learn and grow. He called the package the broadest student-facing AI policy in the country. Some parent groups that had asked for a two-year freeze said the one-year moratorium and the high-school carve-outs fell short.

Teachers still have a back office. Officials said staff may use AI for lesson planning and administration if the tools clear a city safety and privacy review. Every classroom product is supposed to be checked and dropped if it is not mission-critical. A coalition of parents, students, teachers, unions and outside specialists will be asked to revise the rule as the year runs.

The timing is the first week of a new term. Procurement officers now have to strip student logins from chatbot tutors that were sold to schools over the past two years. Vendors who built their pitch on AI literacy in grade 4 lose a customer the size of a small country. Districts that already wrote those tools into curriculum maps have a term to unwind them.

Other large districts have written AI guidance that tells teachers to use with care. New York wrote a ban with a date and a grade cutoff. That is why the announcement travelled. It is also why vendors and some principals will spend the autumn looking for the seams: is a spelling tool that happens to use a language model student-facing generative AI, or is it a dictionary?

The open question is enforcement. A 45-minute cap is a recommendation, not a lock on the iPad. A ban on student-facing chatbots is only as strong as the next teacher who needs a worksheet generated before third period. The city can police official contracts. It cannot police every tab. What it has done is put a written presumption on the table: below ninth grade, the default is no.