Every school in history has tried to ban gum. Every school in history still has gum under the desks. New York City and Los Angeles just restricted student AI on the same day, and if the gum ban is any guide, enforcement is going to be interesting.
On September 2, New York Mayor Zohran Mamdani and Schools Chancellor Kamar Samuels announced a one-year moratorium on what the policy calls student-facing generative AI, education jargon for "the chatbot the kid is actually typing into," from pre-K through 8th grade. That covers nearly 600,000 students, about two-thirds of the system, and companion chatbots are off the table in every grade. Mamdani's pitch was that the tech industry wants everyone to believe AI-powered early education is inevitable, and "We do not see it that way." High schoolers get twice-yearly AI critical-thinking classes and a few tightly supervised pilots, with time limits as tight as 15 to 20 minutes a week for tools like Quill and Edia, which is roughly how long any of us would last before asking the chatbot to just write the whole thing.
Los Angeles got to the same place quietly. The same day, the nation's second-largest district confirmed at a committee meeting that student access to generative AI on district-issued devices had been switched off since the school year began, affecting about 378,000 students. Administrators did it under authority from a June screen-time policy, and it surprised some board members, who heard about it at the meeting. Board member Kelly Gonez summed it up as "no access to things like Claude or ChatGPT," which Gonez called probably the headline. The restriction covers standalone AI services and the AI features baked into Google Workspace.
Here's the loophole. LA's restriction only covers district-owned devices, so a student with a phone and a personal account is, technically, on their own time. It's a gum ban that ends at the school gate. The district also has reason to be cautious: its own AI chatbot, Ed, fell apart in 2024 when the company behind it collapsed, so LA already knows what a bad first date with this technology looks like.
The timing has a punchline. According to Forbes contributor Dan Fitzpatrick, in the days before these announcements, OpenAI expanded free ChatGPT for Teachers to 55 more school systems and Anthropic opened free Claude for Teachers to U.S. districts. New York's policy explicitly lets teachers keep using AI for planning and admin work. Students get the moratorium, teachers get the free samples. And NPR reported this week that research on whether AI actually helps students learn is extremely limited, which means both cities are running a very large experiment with a control group that owns a smartphone.
Both districts have set up committees to study what happens next. If the gum ban is any guide, the data is going to be sticky.
Sources: NYC Mayor's Office — Mayor Mamdani and Chancellor Samuels Put Students First with Nation's Broadest Generative AI Moratorium in Schools (https://www.nyc.gov/mayors-office/news/2026/09/mayor-mamdani-and-chancellor-samuels-put-students-first-with-nat); GovTech — LAUSD Temporarily Bans Generative AI on School-Issued Devices (https://www.govtech.com/education/k-12/lausd-temporarily-bans-generative-ai-on-school-issued-devices); ABC7 Los Angeles — LAUSD Blocks Students From Using AI on District-Issued Devices (https://abc7.com/post/will-artificial-intelligence-banned-lausd-classrooms-new-committee-weigh-rules/19779516/); Forbes — Schools Ban AI For Students While Teachers Get It Free (https://www.forbes.com/sites/danfitzpatrick/2026/09/06/schools-ban-ai-for-students-while-teachers-get-it-free/); NPR via WXXI — Welcome to the 'Wild West' of AI in Schools (https://www.wxxinews.org/npr-news/2026-09-28/welcome-to-the-wild-west-of-ai-in-schools-research-is-scarce-but-experiments-abound)