Two Years to Look

On June 9, a coalition of New York City parents, teachers, and elected officials handed the country's largest school system a letter asking it to stop. They want a two-year moratorium on generative AI in the 1,800 schools that serve roughly 1.1 million children.¹ The petition behind the letter has been gathering signatures since October. It now has more than 3,800. Twenty-nine City Council members, a majority of the body, signed a parallel letter to the mayor and chancellor in support.² Three days after the June 9 delivery, a spokesperson for the Department of Education said the prior administration had "hit the gas on AI without genuine family engagement," and that Chancellor Kamar Samuels would not take that approach.¹

The careful part of the request is not the duration. It is what the two years are being asked to make room for.

What the Pause Is For

A moratorium is, in the plainest reading, a request for time. The coalition behind it, calling itself AIM, has been clear about the question the time is supposed to answer. The signatories include the Parent Coalition for Student Privacy, Class Size Matters, Parents for AI Caution, MORE-UFT, and several others.³ Their letters do not call AI evil. They call it unseen. The objection is not that something dangerous has entered the building. It is that something consequential has entered the building, and the adults who run the building cannot yet see how the children inside it are encountering it. A pause, in that reading, is a place to stand and look.

In late May, after a stormy Panel for Educational Policy meeting and a torrent of comments on the city's March draft guidance, Chancellor Samuels described AI as "the most invasive technology that we've seen" and said the department was looking closely at restricting in-school use for the city's youngest learners.⁴ The instinct is sound. The work is harder. Restriction is a line you can draw on a policy page. Looking is not.

The Line That Has Already Moved

A moratorium can keep AI out of the classroom. It cannot keep it out of the homework. The same parents who signed the petition will, for the next two years, send their children home with assignments. The same children who do not touch a chatbot during the school day will, in many cases, touch one before dinner. The Common Sense census released this month, the first the organization intends to repeat annually, found that eighty-six percent of American kids between nine and seventeen have used a generative AI tool, and that nearly a quarter would ask the tool for help before they would ask a teacher, a counselor, or a parent.⁵ The schoolhouse door is a real boundary. It is no longer the boundary that decides what learning the child is doing.

This is not an argument against the pause. It is an argument for what the pause has to be honest about. Two years of looking will fall short if what the looking surfaces is only the chatbot inside the school. The harder thing the field will have to figure out, with or without a moratorium, is how to make the practice itself visible. The draft a student tried before she opened a tool. The prompt she wrote. The sentence she rejected from the answer. The line she added on her own. Those are the artifacts that distinguish learning from production. They are the artifacts the moratorium, at its best, is reaching toward.

A Future That Holds the Practice

If the next two years go well in New York, the city will not arrive at June 2028 with a final answer about whether AI belongs in school. It will arrive with a clearer set of questions, asked by people who have spent the interval paying attention. The schools that come out of the pause steadier will be the ones that, during it, found a way to watch the moments that matter. Not the eventual essay. The minutes that produced it. The breakthrough at minute thirty-seven that the teacher missed because she was helping someone else. The revision the student abandoned because no one was there to ask her why she walked away from it.

This is one of the quieter reasons Koan has been built the way it has been built. Not as another tool to finish a student's sentence. As a way of holding onto the sentence she wrote, the one she crossed out, and the one that came after. When the practice has a record, the conversation about what role AI should play in it can finally be specific. When it does not, every policy meeting is about a category, and the children inside the category remain, for the adults arguing about them, mostly invisible.

The AIM letter, in its calmest line, asks for time and attention. Those have always been the scarcest goods in school. The question, whether the moratorium is granted or refused, is the same one the field will face anyway. The two years will end. The next two will begin. What will the adults in the building decide is worth looking at?

If your district paused AI tomorrow for two years, what would you spend the interval learning to see?