08/15/2026
It’s our first full school year under Mayor Mamdani, and last week the most expansive student-facing AI policy in the country was rolled out. Students from 2-K to 8th grade are banned from using generative AI software or classroom chatbots. It’s a step in the right direction.
With 1,053 elementary schools and 704 middle schools, the NYC school system is the largest in the entire United States, so quite a few million heads snapped at the announcement on September 2nd that AI is being banned. This system is like the Jamie Dimon of education. Over 900,000 students fall under its purview, so when it speaks, people listen.
The transactional nature of task-completion that characterizes many schools already is prime for the use of AI, which itself is transactional; ask it a question, get a quick and free answer. So a good way to approach this controversial issue is through the lens of learning. All the stuff that happens between question and answer. Does AI help you learn? Yes, but only if you want it to. Jeopardizing learning is what’s at stake here. It’s what needs to be protected from AI. Not doing so would just be overlaying a transformative technology on an educational system that has been largely unchanged since the 18th century–one that doesn't have any guardrails for a student who can offload learning onto a machine built for instant and cogent answers.
But even granting how much is at stake for future generations, there's still a question about whether this is where the mayor's attention belongs right now. Mayor Mamdani ran his mayoral campaign on three pillars: childcare for kids 5 and younger, transportation, and affordable housing. However, we’re 6 months in and he hasn’t advanced on his agenda for k-12 education. In fact, he’s backtracked from ending mayoral control of the school system, reducing class sizes, and boosting pay for paraprofessionals. And this is important because he, in effect, controls the country’s largest school system. It's the largest pillar by both headcount and budget. Amidst issues like declining attendance, enrollment, and a lack of free childcare, tech changes might not be the best use of their time. Largely because unwinding a trillion dollar machine that feeds off our continued use of it will be difficult.
But that difficulty is the point: the harder something is to unwind, the sooner we should start. AI is a highly contested global issue and should be accounted for as such. The market for AI has now surpassed half a trillion dollars, so education’s role in this evolving landscape is under a spotlight because, particularly for young children, the use of AI is like a fever that ought to be caught early. Kids are an unparalleled variation of humans. They possess unique neurochemical processes, high brain plasticity, and a natural drive for exploration. And if you outsource that treasure to a cluster of chips which we already know are capable of learning automatically? We run the very likely risk of producing a less skilled generation. But just complaining about AI won’t solve the problem, an inevitability Mamdani’s administration has conceded to. So about 5% of the public-school student body will be introduced with five limited AI pilots, with built-in limitations and surveillance.
The city school system is also planning to form a coalition of parents, students, teachers, union representatives and experts to help shape the policy. We need as many voices as possible to fight on behalf of our students. The tech industry’s motto is literally to “move fast and break things.” Do we really want our children's intellect and critical thinking skills to be one of those “things”?
Young people already spend most of their lives in the online world, and their habits are influenced by technological developments from an early age. We’ve become largely removed from social interaction because of dense online connectivity. There is a bottomless pit of studies on the adverse effects of that; so what makes us think that letting AI into the classroom will play out any differently? Screenshotting your essay prompt or math homework to ship off to a complex maze of synthetic neurons is not an “innovation.” It’s a setback to our most prized skill: our intellect.
Both in consideration of dissenting views and of the complexity of the issue, an agreeable course of action would be to prioritize social interaction, working memory, selective attention, and cognitive flexibility. All things that generative AI poses a risk to if left unchecked. AI could still have a role here, just not as a shortcut around the actual work of learning. We need to resist simple technological solutions to critical cognitive skills that can be “streamlined” by AI.