New York bans AI in elementary school; Rochester districts choose to teach students to use it without letting it do the work

By Zendoric · September 5, 2026.

New York City has temporarily banned the use of artificial intelligence in its elementary and middle school classrooms, according to News10NBC (WHEC). A few hundred kilometers away, in the districts of the Rochester area, the response is the opposite: not to ban, but to teach how to use it well. Christopher Harris, director of digital learning at Genesee Valley BOCES —the body that coordinates shared educational services among districts in western New York State— sums it up with a figure that no longer surprises anyone: "almost every teenager" has used AI or played with it at some point, in his words.

Harris and his team have received grants to study how AI is used in the classroom, and much of that work consists of something very untechnological: talking to teachers. When a teacher suspects a student has let a chatbot do the assignment for them, Harris's question is not "how do we punish this" but "what are we going to do about it": revisit the exercise's original intent, open a conversation with the student and, if necessary, reconsider whether that was the right assignment.

The other side of the coin is the teachers themselves, who —according to Harris— use AI "much more" than before to prepare their classes; in front of News10NBC's reporters he generated a complete lesson plan in a matter of seconds. But he adds a warning worth underlining: AI "will keep making mistakes", so the teacher needs to be an even stronger expert in their subject —not a weaker one— to spot when the model gets it wrong. That is the underlying paradox of generative AI applied to any expert trade: the more you delegate, the more judgment is required from whoever supervises.

Harris sums up the philosophy with a simple metaphor: using AI in the classroom is like riding an e-bike. The student is still the one holding the handlebars, dodging the potholes and watching the traffic; AI is simply the motor that helps on the climb. "It's a tool for assistance, not for replacement", he says. It is a more honest approach than either an outright ban or unchecked permissiveness: neither of the two solves the underlying problem, which is pedagogical before it is technological.

Our reading. The divergence between New York City and the districts in the west of the state is not just a clash of local policies: it is the same unresolved debate running through education worldwide since generative AI moved into students' homes. Broadly, banning is the easiest response to announce and the hardest to enforce —teenagers already use these tools outside the classroom, with or without the district's permission—, while teaching them to use them requires redesigning assignments, training a teaching staff that has largely received no specific training, and accepting that the line between "help" and "copying" will keep moving.

At Zendoric we have argued, in our coverage of AI and employment sector by sector, that in education the winner is the teacher who orchestrates AI in the classroom —the one who uses it to free up time and reinvest it in the relationship with the student— and the loser is the one who merely transmits content that a model already reproduces just as well or better. This case confirms it in miniature: the time Harris saves generating lesson plans is not a luxury, it is the margin that can then be devoted to the human conversation no chatbot replaces.

In the short term the risk is real and should not be minimized: a generation of students could end up outsourcing critical thinking before having fully developed it, and no district —neither the one that bans nor the one that teaches— yet has solid evidence on which approach better protects that learning in the long run. But the underlying direction points the right way: the task is not to keep students away from a tool they will use for the rest of their professional lives, but to teach them to pedal with assistance without taking their eyes off the road. That is, ultimately, the same skill any profession will demand in the society of abundance that AI promises: knowing when to lean on the machine and when not to.