The Slower Clock

On June 1, Maryland's AI Ready Schools Act, Chapter 634 of the 2026 session, took effect.¹ Governor Wes Moore had signed it weeks earlier as part of the final bill-signing of the General Assembly's session. The legislation is among the most comprehensive state AI laws yet written for K-12 schools. It is also a careful piece of evidence about a question the field has not quite worked out, which is the gap between the architecture of policy and the architecture of looking.

The structure of the law is tidy. The Maryland State Department of Education will publish statewide AI guidance through an online platform, covering data privacy, academic integrity, and age-appropriate use. Each of the state's twenty-four local school systems will have 120 days after that guidance lands to adopt its own AI policy aligned with the state's. Each district will name an AI coordinator. By June 1 of next year, AI literacy will be integrated into the state's computer science and workforce readiness standards. A newly chartered Maryland AI Education Collaborative will sit above it all, studying impact and recommending what comes next.²

It is the most thorough version of a sentence many states are now writing. Ohio's House Bill 96 requires every public, community, and STEM school in the state to adopt a formal AI policy by July 1. Alabama, Idaho, and Oklahoma each have their own laws on the books or about to be.³ By the start of the next school year, more than half the country's students will be in districts under some form of statutory AI requirement. The era of the silence on this question is over.

Two Clocks in the Hallway

The notable thing about the architecture of the new Maryland law is that it runs at one speed and the practice it is meant to govern runs at another. The state writes guidance over several months. Districts have 120 days from that publication to write policy. Coordinators are hired. Curriculum standards are revised by next June. Standards take effect in classrooms the school year after that.

The clock of the statute is set in quarters and academic years. The clock of a seventh grader, deciding whether to ask ChatGPT for help on tonight's history paragraph, is set in seconds. By the time Maryland's standards reach a classroom in the fall of 2027, the eighth grader sitting at her desk will have used a chatbot, more or less daily, for nearly half her life.

This is not a flaw in the Maryland law. It is the inherent timing of policy itself. Statutes set the framework. Students set the pace. The gap between them is where teachers have always done the actual work of teaching, and it is where, this year, much of the question of what AI is doing in school is being decided.

The Look the Law Cannot Take

What the statute will not do, and cannot do, is tell a teacher what a particular ninth grader did with the tool on a particular Tuesday afternoon. It will tell her what her district allows. It will tell her what the academic integrity guidance says. It will tell her there is a coordinator down the hall she can ask. None of that is the look.

The look is the small, attentive act of seeing what happened in the practice itself. The pause before the student typed the prompt. The sentence she rejected from the model. The revision she made after she rejected it. The minute she spent with the problem before she asked for any help at all. The architecture of policy is the easier half. It will be done by the end of next year. The architecture of looking is the harder half, and it is not yet in any state's law. Probably it cannot be. It has to be built by the schools and the classrooms themselves, with tools that make the practice visible without making it surveilled.

A Future That Holds Both Halves

The Maryland law is, by any reasonable measure, the right kind of bill. It sets expectations without dictating implementation. It requires literacy without forbidding use. It creates a body to study impact rather than declaring it. It will improve. The classrooms that thrive over the next decade, though, will be the ones that hold both halves of the question in mind. The policy half, which describes the rules. And the practice half, which is the only place where you can actually tell whether the rules are working.

This is the part Koan was quietly built for. Not the issuance of guidance. The keeping of the record. The trail of how a student used the tool, what she rejected from it, what she added on her own, what she paused over. When that record exists alongside the policy, the AI coordinator down the hall has something to look at when she does her job. When it does not, she is reading a finished essay, the way every teacher has always read a finished essay, with no way to tell whether her district's policy made any difference at all.

The law passed. The clock began. The question now is what the building decides to watch while it waits.

By the time your district's AI policy is finished, what would you wish you had been keeping a record of?