New York City bars younger students from using AI in classrooms and limits teacher grading tools

New York City will prohibit students in grades 2-K through eighth from using artificial intelligence in classrooms starting with the 2026-2027 school year, making it one of the most aggressive school AI crackdowns in the United States. The new policy also stops teachers from using AI to grade student work and bans companion-style chatbots across all grades, signaling a broad effort to slow the spread of automated tools in public education while the city tests a narrow, supervised pilot in high schools.

Mayor Zohran Mamdani announced the changes on September 2, saying the city intends to put human relationships and traditional instruction ahead of rapid adoption of generative AI. The move affects roughly 600,000 public school students and arrives as school districts nationwide struggle with how to balance digital innovation, privacy, and the developmental needs of young children.

The policy does not amount to a total rejection of AI in schools. Instead, it creates a layered system: younger students will be blocked from classroom use, older students will face limited access, teachers will retain some back-office uses, and a small number of high school classes will receive tightly controlled AI pilot tools.

What New York City’s new AI rules do

The core of the plan is a one-year moratorium on AI use for students in 2-K through eighth grade during the 2026-2027 school year. Under the policy, children in those grades will not be allowed to use AI tools in classrooms at all, with one additional restriction barring individual screens until third grade.

For middle schoolers, the Department of Education is also recommending limits on overall screen exposure, including a ceiling of 45 minutes per day. That guidance is paired with a broader school-device policy already set to take effect, reinforcing the city’s push to reduce the role of personal tech in daily learning.

Teachers are not being cut off from every AI use case. Educators may still use the technology to plan lessons, but they will not be allowed to use it to grade assignments. In practical terms, that means AI can assist with preparation, but not with evaluating student work.

Companion chatbots — including systems designed to mimic emotional support or personal friendship — are being banned across all grade levels. That part of the policy reflects growing concern about children forming unhealthy attachments to persuasive AI systems that are not designed to provide real counseling or social development.

Why is New York City restricting AI now?

The city says the policy is intended to protect children’s ability to learn from teachers, collaborate with peers, and work through difficult problems without leaning on software that can shortcut the learning process. Mamdani’s office framed the decision as a response to the idea that schools should slow down rather than rush toward AI adoption.

Children need teachers and human connection in order to learn and grow, Mamdani said in a statement, adding that students need to build relationships with educators, develop alongside classmates and wrestle with difficult problems on their own. He also argued that the tech industry is trying to convince schools that AI-driven early education is unavoidable, and said city leaders reject that premise.

The rationale echoes a wider debate in education: whether generative AI is a useful tutor-like aid, a shortcut that weakens independent thinking, or a privacy and safety risk that schools have not yet fully learned how to manage. New York City is choosing caution, at least for younger students, and is doing so at a scale that makes the decision especially notable.

The district has described the policy as one of the most restrictive in the country. That claim is plausible not simply because of the grade-level ban, but because the city is also limiting how teachers may use AI and closing the door on companion bots altogether.

How will high school students be allowed to use AI?

High school students will not get unrestricted access. Instead, they will be allowed to use AI only in specific situations, and the city says they will also receive AI literacy lessons twice a year to help them evaluate the technology critically before becoming dependent on it.

The literacy component matters because the policy is not only about prohibition. It is also about instruction. By pairing limited access with recurring lessons, the city is trying to teach students how to question outputs, spot errors, and understand when AI should and should not be used.

That approach reflects a growing consensus among educators that older students need to learn about AI whether or not schools like it. Even district leaders who distrust broad deployment increasingly acknowledge that students will encounter the tools in college, the workplace, and daily life. New York’s answer is to delay use for younger grades and introduce it cautiously later.

What the city says students will learn

The twice-yearly sessions are meant to help students think critically before relying on AI. In practice, that likely means lessons on prompt quality, hallucinations, bias, privacy, and the difference between assistance and plagiarism. The city has not framed the classes as a coding initiative; rather, they are a digital judgment and awareness exercise.

That distinction is important. Many school districts have focused on teaching students how to build with AI. New York is emphasizing how to evaluate and control it.

Which AI tools are included in the pilot program?

The city is not banning AI outright in every classroom experiment. It will run a limited pilot in up to five classes in a small number of high school classrooms, using tools that have been screened for safety and data privacy and that keep students as the “primary thinker.”

The pilot includes several named products:

• Quill for English language support
• Edia for math
• Brisk Teaching for guided study practice
• Playlab and Intel AI-Ready Schools for support across multiple subjects

These tools come with time limits and are meant to be used in a controlled setting. The city’s description suggests a model where AI acts as a constrained helper rather than an autonomous tutor. That is consistent with the broader policy goal of keeping students in charge of the learning process.

How the pilot differs from everyday classroom AI

The pilot is intended to be narrow, supervised, and measurable. Unlike open-ended use of consumer AI platforms, the program only includes vetted tools with privacy standards in place and usage limits built in.

That matters because one of the biggest concerns in schools has been data exposure. Consumer AI products often collect prompts, retain chat histories, or rely on uncertain training and retention practices. By contrast, a vetted pilot can be evaluated against school-specific rules for security and student protection.

How does this compare with other school technology limits?

New York City’s AI restrictions follow closely on the heels of a separate rule limiting personal devices. The public school system already introduced a bell-to-bell ban on personal internet-connected electronics, including smartphones, for the 2025-2026 school year. Together, the two policies amount to a stronger retreat from constant digital access than many districts have attempted.

That broader device strategy helps explain the AI decision. Once schools restrict phones and tablets, the argument for unfettered classroom AI becomes harder to make, especially for younger students who are still building foundational reading, writing, and interpersonal skills.

At the same time, the city is not alone in trying to rein in technology in education. Officials point to other countries, including Norway, Sweden, and the Netherlands, where governments have moved to address concerns about how digital tools affect children’s attention, development, and learning habits. The specifics differ, but the underlying impulse is similar: schools should not adopt every new platform simply because it exists.

Why this policy may matter beyond New York

Because New York City runs one of the largest public school systems in the United States, its decisions can shape national debate. A policy affecting 600,000 students carries symbolic weight far beyond the five boroughs, especially when it sets limits that many districts may be reluctant to impose.

If the rules hold up in practice, other systems could use them as a template for age-based AI limits, teacher-use restrictions, and curated pilot programs. If the policy is seen as too rigid, it may instead become a cautionary example for districts weighing how much control to place on emerging tools.

What educators and parents should watch next

The biggest questions now are implementation and enforcement. Policies can be announced quickly, but classroom habits change more slowly. The city will need to define what counts as AI use, how exceptions are requested, and how schools verify whether a tool is acceptable under the new rules.

Parents will also want clarity on what happens outside classroom hours. If students are prohibited from using AI in school, some may simply shift the behavior to home. That raises a familiar challenge for school policy: controlling the environment inside school is possible, but controlling the broader digital ecosystem is much harder.

Another unresolved issue is whether the pilot will influence the city’s future stance. A successful small-scale trial might eventually expand AI use in selected grades or subjects. A failed or controversial trial could strengthen the case for longer restrictions.

Key implementation questions

1. How will schools define and detect AI use across different software platforms?
2. What exceptions will be granted for students with disabilities or multilingual learners?
3. Will the city expand the high school pilot if the tools prove effective?
4. How will teachers be trained to follow the grading and lesson-planning rules?

Those details may prove just as important as the headline ban. In school systems, the difference between a symbolic policy and a workable one often comes down to training, oversight, and parent communication.

What exceptions are built into the ban?

The policy does include exceptions, and those carveouts reveal where the city sees legitimate educational value in AI. Students with disabilities may be allowed to use approved tools, and multilingual students may also receive exceptions when AI helps with access or language support.

Schools may additionally approve specific tools for programs such as computer science. That leaves room for AI to be studied as a subject or used in technical courses, even while ordinary classroom use is restricted.

These exceptions are important because they keep the policy from becoming purely symbolic. The city appears to be distinguishing between assistive, educationally necessary uses and broader everyday dependence on generative systems.

How are schools trying to balance innovation and caution?

They are doing it by separating use cases into categories. Younger students face prohibition, older students face limited access, teachers can use AI for planning but not grading, and a small pilot tests vetted tools under supervision.

That structure suggests the city is not trying to make a philosophical statement that AI is inherently harmful. Instead, it is drawing a developmental line: younger children should learn without it, high school students should learn about it carefully, and only selected tools should be allowed into classrooms under formal review.

In that sense, New York’s policy reflects a broader moment in education policy. Schools are no longer asking simply whether AI should be allowed. They are asking who should use it, at what age, for what purpose, with what safeguards, and under whose supervision. The city’s answer is to start with restraint.

Timeline of the new NYC school AI policy

The changes arrive in stages, not all at once. That gradual rollout gives schools time to prepare and families time to adjust.

The bigger picture for AI in education

New York City’s move underscores a central tension in education technology: schools want the benefits of AI without importing its risks. But the more powerful and accessible the tools become, the harder it is to keep them from reshaping how students complete work, how teachers manage time, and how institutions define learning itself.

For supporters of the new policy, the answer is simple: children should master reading, writing, reasoning, and social interaction before they hand those tasks to software. For critics, the risk is that blanket restrictions could leave students less prepared for a future where AI is already embedded in the workplace and higher education.

The city’s policy attempts to split that difference by delaying widespread adoption rather than rejecting the technology entirely. Whether that compromise succeeds will depend on how well the restrictions are enforced and how effectively the high school pilot and literacy lessons prepare students for a world where AI is unavoidable, even if it is not welcome in every classroom.

For now, New York City has chosen a conservative path: more human instruction, fewer screens, no AI companions, and only tightly controlled experimentation. In the nation’s largest school system, that may be the clearest sign yet that the educational backlash against rapid AI adoption has moved from debate to policy.