Summary
On September 2, 2026, New York City Mayor Zohran Mamdani and Schools Chancellor Kamar Samuels announced a one-year moratorium on student-facing generative AI for all public school students from 2-K through eighth grade, covering nearly 600,000 students, about two-thirds of the district's enrollment, effective for the 2026-27 school year. The policy bans companion chatbots in every grade, disables the AI components of more than 38 previously approved classroom products, caps high school AI use to five vetted pilots for at most 50,000 students, and introduces recommended daily screen-time limits by grade. Later the same day, the Los Angeles Unified School District, the second-largest US district, barred all students from using generative AI tools on district devices for the 2026-27 school year. The two decisions reversed the course both districts set in 2023 and 2024, when New York lifted an early ChatGPT ban and Los Angeles commissioned its own student chatbot.
What Happened
The New York policy, announced at the Brooklyn STEAM Center, applies to "all software that uses student-facing generative AI" for students in 2-K through eighth grade, with exceptions for assistive technology used by students with disabilities, multilingual learners and students in career-readiness programs such as computer science. Teachers may continue to use AI for planning and operational tasks if the tools meet the district's safety standards. Companion chatbots are prohibited in all grades. In his remarks, Mamdani said the district would "fully discontinue or disable the AI components of more than 38 previously allowed programs that do not meet our new safety and oversight standards." Asked whether that covered citywide contracts such as the Amira reading tutor, Google Gemini on student Chromebooks and HMH curriculum products, senior education adviser Ailish Brady said AI features embedded in Tier 1 curriculum "will be turned off," including Amira and HMH products, and that any product whose AI cannot be disabled "will be taken off the board entirely."
For high schools, the district authorized five pilots, capped at five classes per school and 50,000 students in general-education classes, about 5% of enrollment: Quill for English language arts, limited to 15 minutes per week; Edia for mathematics, limited to 20 minutes; Brisk Teaching for teacher-created activities, 10 to 20 minutes once or twice a week; Playlab, no more than two assignments per marking period; and Intel AI-Ready Schools, one period per week for semester-long projects. All pilot use is under direct supervision of a trained educator. Mamdani said that "none of these programs are chatbots" and that the pilot "is not introducing ChatGPT, Claude or anything similar into our curricula." All high school students will take two 45-minute AI literacy modules a year covering what AI is and is not, bias, risk, ethics and effects on careers.
The accompanying screen-time policy restricts one-to-one device time for students in second grade and below and recommends caps of 30 minutes a day for grades three through five and 45 minutes for grades six through eight. The district said every technology tool in use would undergo a review covering vendor accountability, evidence of learning design, prior research and user feedback, and would be barred if not "mission-critical to learning." A Technology in Schools Coalition of students, educators, parents, elected officials, advocates, union representatives and outside experts will assess the moratorium and the pilots during 2026-27 and publish recommendations for subsequent years. The press release carried supporting statements from Governor Kathy Hochul, Public Advocate Jumaane Williams and several state legislators and council members, including Assemblymember Robert Carroll, who said he had introduced similar legislation at the state level.
Los Angeles Unified announced its restriction at a committee hearing on the same day. Bloomberg and The Washington Post reported that the district barred all students from generative AI tools on district-issued devices for the 2026-27 school year while a new Generative AI Ad Hoc Committee develops recommendations for the school board by the end of the year; the committee held its first meeting that day. Before the change, Los Angeles students aged 13 and over who had completed digital-citizenship instruction could use approved generative AI tools. The district's board had adopted a grade-by-grade screen-time policy in June 2026.
Both districts had previously moved in the opposite direction. New York's Department of Education blocked ChatGPT on school devices and networks in January 2023 and then dropped the ban in May 2023, when then-Chancellor David Banks wrote that students "will work in a world where understanding generative AI is crucial." Los Angeles under former superintendent Alberto Carvalho, who resigned in June 2026, had commissioned a district chatbot for students. Axios reported on September 7 that Chicago school board candidates were signing a pledge for a three-year moratorium on generative AI in that district, and cited an NPR/Ipsos poll in which 55% of teachers said students mostly use AI as a shortcut and more than half said their school had given them no guidance on it.
Why It Matters
This is the first time the two largest US school districts have removed generative AI from ordinary instruction after having adopted it, and New York's version is a written policy with named products, time limits and a review body rather than a device-level block. The list of disabled products shows how far generative features had already spread into standard curriculum vendors and into Google's Chromebook software, which is why the district framed the decision as a moratorium plus procurement standards rather than a ban on a single app. Together with the companion-chatbot prohibition in all grades, the policy treats student-facing AI as a category needing evidence before use, a position closer to the district's January 2023 stance than to the guidance it published as recently as March 2026, which a reporter at the briefing noted had listed brainstorming, organizing, refining communications and translation as acceptable uses.
The moratorium is explicitly temporary and the pilots are small, so its durability depends on what the coalition reports in 2027 and on whether state legislation such as Carroll's advances. Enforcement is an open question: Axios noted that Gemini can appear in searches on school Chromebooks, and the district has not published the list of affected products or how it will verify that vendors have disabled AI features. No evaluation design for the pilots has been released, so the year of study the mayor described may not yield comparable data. Whether other large districts follow, and whether vendors respond by shipping AI-free variants for K-8 markets, will determine if this becomes a national pattern or a New York and Los Angeles exception.
§ How to read the metadata
- Landmark
- Fundamentally alters the trajectory; 2–5 per year.
- Major
- Meaningfully shifts the landscape; 2–4 per month.
- Notable
- Worth documenting; significance can be upgraded later.
- Confidence
- High = primary sources corroborate. Medium = credible secondary only. Low = provisional. Disputed = credible sources disagree.
- Contestation
- Uncontested = no formal challenge. Contested = at least one challenge open. Superseded = replaced by a later entry. Unresolved = dispute still open.