New York Banned AI in Schools. Boston Made It a Graduation Requirement.

New York City just banned AI for students through eighth grade. Los Angeles barred it for everyone. The same week, Boston Public Schools started rolling out a requirement that every high schooler graduate AI fluent, the first major American city to make it mandatory. Three districts, three headlines, three opposite answers to the same question: what do we do about AI in schools?

If you run a district and you are watching this unfold, you already know the real question is not ban or require. It is which lever actually protects kids and prepares them, and almost nobody is asking that question in public. They are picking a side and calling it a policy.

What New York and Los Angeles actually did

New York City, the nation’s largest school district, opened the school year with a one-year moratorium on generative AI for its youngest students and tighter screen time limits across the board. Officials said they would disable the AI components in more than three dozen previously approved programs that did not meet new safety standards. High schoolers get twice-yearly critical thinking modules and a small number of supervised pilots. Everyone else waits.

Los Angeles Unified, the second largest district in the country, went further. Students who were allowed limited AI access last year under certain conditions lost that access this year, while the district’s generative AI committee works out a curriculum. Both moves came from the same instinct: the tools moved faster than the safeguards, so pull back until the safeguards catch up.

What Boston actually required

Boston Public Schools took the opposite bet. Starting this September, AI fluency becomes part of what it takes to graduate from a Boston high school. The rollout includes teacher training, an industry built curriculum developed with UMass Boston, student hackathons, and career pathways, backed by a public private partnership the city announced earlier this year. The premise: students who leave high school without knowing how to use AI, question it, and catch its mistakes are walking into a workforce that already expects that fluency. Waiting is its own risk.

Neither city is wrong. That is what makes this useful instead of just newsy.

Why the ban versus mandate framing misses the point

A ban answers a control question. A mandate answers a readiness question. Those are different problems, and a lot of district leaders are treating them as the same decision because the headlines arrived in the same week.

Here is the invisible part neither approach fixes by itself: a signed policy, whichever direction it points, does not tell a single teacher what to do on a Tuesday. It does not tell a school business office which vendor tools are safe to keep. It does not tell a counselor what to say when a parent asks if the ban applies to the AI tutor the district already pays for. The policy is visible. The daily judgment calls that make the policy real are not, and that is where most rollouts actually succeed or fail.

Should your district ban AI or make it a graduation requirement?

Probably neither, as a first move. Before you pick a stance, answer three questions honestly.

First, what problem are you actually solving? If the concern is younger students outsourcing thinking to a chatbot, a ban addresses that directly. If the concern is graduates arriving at a job or a trade program without basic AI fluency, a mandate addresses that directly. Most districts are worried about both at once, which is exactly why a single blanket policy tends to satisfy no one.

Second, who in the building can actually explain the policy the way it was written? Ohio, Idaho, Maryland, Oklahoma, and Virginia already require districts to have an AI policy on paper. Having one and having a staff that can explain it under pressure are not the same accomplishment. A ban nobody can articulate is not meaningfully different from having no policy.

Third, have the adults been trained, regardless of which way the district leans? This is the part almost every district skips. Boston is investing in training alongside its mandate. New York and Los Angeles have not said much yet about training the teachers who now have to enforce a ban inside classrooms full of students who already have AI on their phones. A rule without trained adults behind it is a document, not a practice.

What this means for the business office, not just the classroom

A ban or a mandate reads like a curriculum decision. It is not. Every AI tool already sitting in your stack, the writing assistant in the learning management system, the transcription tool in special education meetings, the chatbot vendors keep pitching to your procurement office, now has to be checked against whichever direction your district picked. A ban that only covers classroom instruction and never touches the tools your operations team already licensed is not a ban. A mandate that trains students but never audits what the adults are quietly already using is not a mandate either.

This is where school business officials do the work nobody sees. Before you announce anything, pull the list of every AI-enabled tool currently under contract, ask each vendor how it uses student data, and flag anything that contradicts the stance you are about to take publicly. That audit takes a few hours. Skipping it and finding out about a contradiction from a parent’s public records request takes a lot longer, and it undoes trust you cannot easily rebuild.

The part that actually determines what happens next

Picture two districts a year from now. One banned AI and never built the muscle to revisit the decision as the tools change. The other mandated AI fluency and never trained teachers well enough to teach it responsibly. Both end up in the same place: a policy on file and a staff improvising in the classroom. The headline was the easy part. The follow through was always the real work, and it was always going to be invisible from the school board table.

If your district is choosing a side this fall, choose the side that matches the problem you are actually trying to solve, then fund the part nobody photographs: training the people who have to make the policy real every day, not just the policy itself.

That is the work worth doing before the next headline forces your hand. If your leadership team needs a working session to build that plan, not another slide deck, training is where that starts. If your board or your staff needs the case made in the room, in person, a keynote is the place to make it land.