New York City Bars Classroom AI Through Middle School for One Year

What's being blocked are generative AI learning tools that students use directly — chatbots that answer questions, write text for them, or generate problems and solutions. The pause takes effect for the 2026-27 school year, and the announcement lands just as New York City public schools are opening the new academic year.

This is less an outright ban than a one-year moratorium. During that time, the education department will review every technology tool used in schools, evaluating vendor accountability, safety, transparency standards, ethics, classroom impact, supporting research, and user feedback.

What's actually being switched off is a list of tools. Generative AI features are being disabled in roughly 38 edtech products, including the reading-support tool Amira Learning, the HMH product line, and Google Gemini.

Screen Time Limits Grow With Grade Level

The district also drew lines around devices themselves. For pre-K through second grade, one-to-one distribution of school-issued devices is restricted. For third through fifth grade, the recommended daily screen-time cap is 30 minutes; for sixth through eighth grade, it's 45 minutes.

These caps apply to overall screen time, not just AI use — and they're recommendations, not mandates.

It echoes an old question about when to hand a child a calculator. Give kids a calculator before they've memorized their times tables, and arithmetic gets faster — but the chance to actually learn arithmetic disappears. That's the logic behind blocking the entire elementary-through-middle-school range: the goal is to keep AI from doing the work during the years when the muscles for reading, writing, and math are supposed to form.

Exceptions are clearly carved out. Students with disabilities who need assistive technology, multilingual learners, and centrally approved curriculum programs — including coding and computer science — remain unaffected.

Companion Chatbots Banned at Every Grade

The toughest provision applies across the board. Companion chatbots — conversational AI positioned as offering emotional or psychological support — are banned for every grade level, including high school. These are AI systems designed to act as an ongoing friend or counselor, building a continuing relationship with the user.

While learning tools face a one-year pause, companion chatbots are blocked outright, with no exceptions. That structure leaves room for this particular rule to remain in place even after the moratorium on other tools ends.

Teachers are treated differently. As long as a tool passes the education department's safety standards, teachers can keep using AI for lesson planning and administrative work. In effect, the policy separates AI as a work tool for adults from AI as a learning partner for children.

High Schools Get Five Pilot Paths

High schools aren't shut down entirely. Pilot programs will run with five vetted tools — Quill, Eddia, Brisk, PlayLab, and Intel AI Ready Schools — capped at five classrooms per school.

The education department said the pilots follow an evidence-based approach and are focused on confirming whether student achievement actually improves. The Wall Street Journal reported that an AI literacy course for high school students is also being offered alongside the pilots.

Schools Chancellor Karma Samuels said, "We need to prepare our older students thoughtfully and deliberately for a world where AI is already part of everyday life."

'We Don't Believe It's Inevitable'

The most-quoted line from Mayor Mamdani's announcement takes direct aim at the tech industry: "The tech industry wants us to believe that AI-driven early education isn't just inevitable but necessary. We don't see it that way."

Pushback came immediately. Jean-Claude Brizard, president of Digital Promise, raised concerns about equity — children from well-off families will have plenty of access to AI outside school, while it's low-income kids who'll effectively lose access.

Enforceability was also questioned. Kelly Clancy of Parents for AI Caution pointed out that the screen-time limits are only recommendations with no binding force, and Keri Rodrigues of the National Parents Union criticized the process as a "political mess," saying parents weren't adequately consulted.

New York City isn't alone in this. Los Angeles, the second-largest school district in the country, has also restricted screen use, and similar bills have emerged in multiple states across the political spectrum.

Editor's Take

Reading this announcement as a Luddite backlash misses the point. The real target of the ban isn't students — it's unvetted vendors. Teachers keep using AI, coding classes continue as before, and high schools get five pilot programs. What's actually being blocked is a situation where products that never passed safety standards had become the default in 600,000 students' classrooms.

That's especially true of the detail about switching off generative features in 38 products. The education department isn't returning something it just bought — it's rewinding a situation where tools already in use quietly turned on AI features through updates. Features that didn't exist when schools signed their contracts had simply appeared on kids' screens one day. The one-year pause is time to reclaim that control.

It's also worth noting that companion chatbots are the only category banned without any grade-level distinction. There's room to debate whether learning tools are effective, but no such room was given to chatbots that set out to handle children's emotional needs. Given everything that's happened around teens and AI companion apps over the past year, this is likely the provision with the most staying power.

It's clear why Korean education circles will cite this case. Korea is working through the same questions with its AI digital textbooks, but the debate there has mostly split into two camps — use it or don't. New York City has placed a third option in between: cutting the question along grade level, along use case, and along vendor vetting.

The equity critique carries real weight. Even if schools block AI, kids who use it at home will keep using it, which could widen the gap rather than close it. But for that critique to hold, you first have to assume that today's classroom AI actually helps learning — and confirming that assumption is exactly what New York City says it plans to spend the next year doing.

A decision by the largest school district in the country becomes the baseline for the next contracting cycle. Edtech companies will now need to knock on school doors with "we passed this safety standard," not "we added AI." Whatever list New York City produces a year from now, the rubric behind it is likely to get copied into procurement documents for school districts across the country.