What A Framework Asks That A Ban Cannot

On Monday night, two days before the first day of school, the Shawnee Mission School District in Kansas presented its new classroom technology plan to the public for the first time.¹ The plan was not the one most people expected. In July, board discussions had leaned toward a total ban on generative AI on district-issued devices.² Parents from a national coalition called Schools Beyond Screens had shown up in numbers, wearing buttons, asking for a moratorium on AI and less screen time in general.³ The district heard them. Then, on Monday, it did something more interesting than a ban. It named a framework.

The framework is called PICRAT.¹

Two Axes, Nine Cells

PICRAT is a research-based teaching model that asks two questions about any use of technology in a classroom.⁴ The first question is about the student. Is the student passive with the tool, or interactive with it, or creative with it. The second question is about the teacher. Is the tool replacing something the teacher already does, or amplifying it, or transforming it. Nine cells in a small grid, ordered from a bottom-left corner labeled passive replacement to a top-right corner labeled creative transformation. A first grader watching a video on a Chromebook sits in one cell. A fifth grader using AI to test her own argument sits in another. A twelfth grader publishing a documentary that could not have existed a year ago sits in a third.

Nobody in the framework is arguing that every use of technology has to reach the top-right corner. What the grid asks the teacher to do, before any tool is opened, is name where on the grid the lesson actually is. That is a quieter demand than a policy. It is also a harder one to skip.

The Third Answer

Most districts this summer are choosing between two answers about generative AI. Ban it. Or allow it, with training. Shawnee Mission tried, briefly, to choose the first one, and its parents did not quite let it. Then it chose a third answer that is neither. The district's older students will be able to access built-in AI tools on their devices, but only under direct teacher guidance. The younger grades will have strict limits.¹ A broader, permanent AI policy is still being drafted, and the district plans to form a community workgroup to help write it.² In the meantime, teachers will decide, one lesson at a time, whether the tool in front of the child is asking her to be passive, or interactive, or creative.

The distinction the district is making, without quite saying it, is that the object of a policy should not be the software. It should be the encounter between a child and the software. A chatbot used to summarize a reading is one encounter. A chatbot used to test the strongest counterargument to a fifth grader's own paper is another. The tool is the same. The encounter is not. A framework can hold that distinction. A ban cannot.⁵

The Future That Reads Like This

The most useful thing about PICRAT, if the country decides to borrow it, is not the grid. It is the habit the grid enforces. A teacher who has to place a lesson in a cell has to look, briefly, at what the child is actually being asked to do with the tool. That look is the thing most AI policy debates have skipped. When the look becomes routine, a lot of arguments that feel unresolvable resolve on their own. A tool that keeps the child passive is one story. A tool that helps her produce something she could not have produced alone is another. The right policy is different for each.

The future of school technology, if it is a good one, will look like a thousand of these small teacherly looks, repeated across a country of classrooms, aggregated into something like an evolving common sense.

The Layer The Grid Points To

At Koan we think about the layer of a child's work that a tool touches. The PICRAT grid, without meaning to, points at that layer. When a student is passive with a tool, the layer is thin. There is not much of her work to see. When she is creative with a tool, the layer is thick. Her thinking, her revisions, her pauses, her breakthroughs are all in it, and they can be read afterward by a teacher who wants to know what happened in the child's head this week. That thicker layer is what a school protects when it takes the framework seriously. The grid is only useful if the record of the child's thinking, on the other side of the encounter, is legible enough to look at.

Shawnee Mission did not name that second half of the problem yet. Most districts have not. But naming the first half is more than most districts did this summer, and it is the harder half to name.

If your district could ask, every Monday morning, one question about every tool it lets into a classroom, would that question be about the brand of the tool, or about the child holding it?