New York City bans generative AI for 600,000 students through eighth grade

New York City will bar nearly 600,000 public school students from pre-kindergarten through eighth grade from using generative AI tools during the 2026-27 school year, in what Mayor Zohran Mamdani's office describes as the most expansive prohibition on student-facing AI in the United States.

The one-year moratorium, announced on 2 September together with a new AI and Screen Time Guidance, covers all software using student-facing generative AI - chatbots, AI tutors and AI-powered instructional programs - for pupils in grades 2K through 8. That group represents roughly two-thirds of enrollment in NYC Public Schools, the largest public school system in the country. Companion chatbots are prohibited across every grade, high school included.

High schoolers face a narrower regime. All students in grades 9 to 12 will complete two 45-minute AI literacy modules each year, covering how AI works, data privacy, bias and appropriate use in school. A limited set of five vetted pilot programmes may reach a maximum of 50,000 high school students, about 5% of the district, with no more than five classes per school and direct supervision by a trained teacher. The pilots are Quill for English language arts, Edia for math, Brisk Teaching, Playlab and Intel AI-Ready Schools, each carrying weekly time caps of roughly 15 to 45 minutes. General-purpose chatbots such as ChatGPT are not part of any pilot.

Teachers keep considerably more room to manoeuvre. They may use approved AI tools for lesson planning, translation and drafting family communications, but not for grading student work, monitoring student behaviour, writing individualised education plans or supporting students in mental-health crisis.

The guidance also introduces grade-based screen limits: no individual devices for pupils in 2K through grade 2, a recommended maximum of 30 minutes a day for grades 3 to 5, and 45 minutes for grades 6 to 8. Assistive technology, assessments, remote-learning days and centrally approved instruction are exempt, as is technology supporting students with disabilities and English language learners.

On procurement, the city says it will disable the AI components of 38 previously approved education technology contracts that do not meet its new safety and oversight standards; the vendor list has not been published. Officials singled out digital tutoring software Amira and Houghton Mifflin Harcourt's Into Reading digital reading curriculum as products whose AI functions will be switched off. Amira said it has contacted the education department to clarify what the policy means for its product; HMH said its elementary reading course contains no student-facing AI features.

The reversal is sharp. In January 2023 the district blocked ChatGPT on school networks, then lifted the block four months later, with then-chancellor David Banks calling the initial restriction a "knee-jerk fear". Three years, a new mayor and a new chancellor later, the city has gone further than it did in 2023.

During 2026-27 a Technology in Schools Coalition of students, educators, parents, officials, union partners and outside experts will assess the impact of the moratorium and the high school pilots, and publish recommendations for later school years. Schools chancellor Kamar Samuels has acknowledged that an earlier AI framework issued in spring 2026 "missed the mark" with parents and teachers. Open questions remain over how the city will verify that AI features have actually been removed from vendor products, how compliance will be enforced, and how out-of-school use will be treated.