The Chair They Put A Student In

Last Wednesday, at the Bridges public sector AI summit in New York City, a student sat on a panel of school district leaders and described what it felt like to be handed a seat at the policy table.¹ "Together we all drafted the district AI policy," she said, "and I, a student, was actually sitting there at the table typing into that very document that became the AI policy that we now use at my school."¹ Then she said the sentence worth pausing on. "It's not enough to just handpick three students that can be sitting at that table. There should be a system that invites students to be a part of that agency."¹

The word doing the interesting work is system.

What The Panel Was Arguing

The panel's headline argument was that AI decisions should start with students, not be made for them.¹ Speakers named the alternative bluntly. Top down. Handpicked. Ceremonial. If schools want technology to meaningfully improve teaching and learning, they said, students must become active participants in shaping how it is used, as co-leaders of the work rather than an audience for it.¹ A companion session went further, arguing that the most effective AI strategies are less about choosing the right technology than about building trust, defining instructional goals, and creating systems that continually evaluate whether AI is actually improving learning.²

Any district leader who has watched a student advisory committee run for a year knows the temptation to stop short of that. Choose three articulate seniors. Sit them next to the assistant superintendent. Print the roster. Call the input real because a student's name is on the page. The Bridges panel was quietly refusing that version. What it did not have time to answer, on stage, was the harder question. What is a policy conversation with a student actually supposed to be about?

The Gap Under The Chair

The number worth putting next to the student's quote is one that landed earlier this month, from the Digital Education Council's global survey of forty five thousand higher education students and faculty across thirty five countries.³ Eighty eight percent of students say they use AI in their learning.³ Twenty nine percent of students believe their instructors are equipped to guide them through it.³ The gap between those two numbers is not, strictly, a policy gap. It is an evidence gap. The adults in the room can name the tools. They can quote the terms of service. What they cannot do, in most classrooms, is describe what a specific student actually did during the twenty minutes between opening the document and turning it in.

You can put a student in the chair at the policy table. You can hand her a laptop. You can let her type sentences into the draft. If nobody else in the room can see what happened during her last essay, though, the sentences she types will be, at best, an anecdote. At worst, they will be a performance. Either way, the policy will still be written by the people with the loudest guesses.

What A System Would Need

The student at Bridges asked for a system. She used the word precisely. Not a committee. Not a panel. A structure. A structure that a hundred students in a district could plug into, not three. A structure that would still work in a middle school where nobody wants to speak in a room of adults. A structure where the input was not a testimony but a trace.

That is a demanding request. It has almost nothing in common with the way districts currently think about student voice. It also has almost nothing in common with the way tech companies currently sell AI tools into schools. It has, though, a lot in common with the way a good teacher, sitting next to a student for an afternoon, learns what a piece of software is really doing to the way that student thinks.

The Layer We Are Building

The layer we build at Koan is that trace. When a student writes, she pauses. She deletes and rewrites. She copies something in, sometimes leaves it, sometimes translates it into her own language. None of that shows up in the finished paragraph. All of it is exactly what a policy committee would need to see, from a wide sample of real students, to know whether a tool is thinning cognition or thickening it. Making learning visible is not, in the end, only about grading. It is the substrate under any honest conversation about which tools to let in.

A student at a policy table is a strong signal about intent. A record of what students actually did last week, across a school, is what turns that signal into a decision.

What The Fall Will Actually Decide

By fall, most of the districts still writing AI policy this summer will publish some version of a document that names student voice as a principle. A few will have done what the Bridges panel argued for. Most will have handpicked three students and printed the roster. The distance between the two, in practice, is not the number of students in the chairs. It is the quality of the evidence any of them can point to when the conversation gets specific. A policy written from finished essays is a policy written from finished essays. It will be well meaning. It will not be enough.

If the students at your policy table can only describe what AI does to their work in the abstract, what is your policy going to be based on?