Series Overview
Designing with Friction: AI, Learning, and K–12 Leadership
Generative AI is changing more than the tools students and educators use. It is changing the conditions under which work is produced, learning is demonstrated, data is remembered, and institutional decisions are made. This five-part series examines what K–12 leaders need to preserve, redesign, and govern as AI becomes embedded in classrooms, platforms, procurement, policy, and school operations.
The central argument is that schools should not respond to AI only through bans, detection tools, or broad acceptable-use language. They need a clearer framework for protecting the human work of learning: thinking, explaining, revising, questioning, judging, and taking responsibility for ideas. That requires distinguishing productive friction, which supports learning, from exclusionary friction, which blocks access without educational value.
In this series:
Part 1: The Illusion of Completion examines why polished student work may no longer be reliable evidence of learning.
Part 2: Writing Policy for the Learning Sequence argues that AI policy should address when AI enters the learning process, not only whether a tool is allowed.
Part 3: From Assistant to Actor explores why agentic AI creates new procurement and governance challenges for school districts.
Part 4: The Compliance Ghost in the Machine focuses on persistent AI memory, student data privacy, and the risks of invisible personalization.
Part 5: Tiered Decision Rights offers a framework for deciding what AI may generate, recommend, automate, and what must remain human.
You are reading Part 2: This post shifts from tool permission to learning sequence, asking when AI should enter the task.
Writing Policy for the Learning Sequence: Moving Beyond Tool Permission
Most school AI policies begin with a simple question: Is this tool allowed?
That question matters. Districts have real obligations related to privacy, security, accessibility, bias, procurement, and age-appropriate use. Schools cannot responsibly ignore vendor terms, student data practices, or compliance requirements. But tool permission is only the first layer of AI governance.
The deeper instructional question is different.
When should AI be allowed to enter the learning process?
That question matters because the same tool can either support learning or bypass it, depending on where it appears in the sequence of the task.
Consider a student writing an argumentative essay. If the student uses AI before reading the source material, before developing a position, and before struggling to organize evidence, the tool may complete the most important cognitive work before the student has begun. The student may end with a polished essay, but the learning sequence has been hollowed out.
Now consider a different sequence. The student reads the sources first. They annotate evidence, draft an initial claim, identify uncertainty, write a rough argument, and then use AI to test whether the reasoning is clear. In this case, AI may function as a scaffold. It can provide feedback, identify weak transitions, suggest counterarguments, or help the student notice gaps. The technology has not replaced the thinking. It has entered after the student has done enough work to judge the output.
That distinction is central to responsible AI use in schools.
Too often, AI policy is written as if the main issue is permission. Students may use AI for brainstorming but not drafting. Teachers may use AI for planning but not grading. Staff may use approved tools but not unapproved tools. These rules are useful, but they are incomplete because they do not address the order of learning.
A stronger policy framework should include sequence-based guidance.
A sequence-based AI policy asks which parts of a learning task must remain human first. It does not assume that AI use is always harmful or always helpful. Instead, it identifies the points in the learning process where students need productive friction before automation is introduced.
A simple district policy principle could read as follows:
Students must complete designated thinking, drafting, problem-solving, reading, evidence-gathering, or explanation steps before using AI tools for feedback, revision, comparison, or extension, unless the teacher has intentionally designed the task otherwise.
This kind of language does several important things.
First, it keeps the teacher in the role of instructional designer. The teacher decides when AI supports the learning goal and when it undermines it. This is more realistic than asking teachers to police every possible use after the fact.
Second, it separates access support from cognitive bypass. A student using AI translation to understand directions may be reducing exclusionary friction. A student using AI to generate an entire response before engaging the content may be bypassing productive friction. Policy should help educators make that distinction rather than flattening all AI use into “allowed” or “not allowed.”
Third, it creates a shared language across classrooms. Without sequence guidance, AI expectations vary wildly from teacher to teacher. One classroom may treat AI as cheating. Another may treat it as a normal writing partner. Another may avoid the question altogether. Students are left guessing, and families receive mixed messages. Districts need enough coherence that students understand not only whether AI may be used, but when and why.
Fourth, it shifts the conversation away from detection. AI detection tools are unreliable as a foundation for instructional policy. A district cannot build a learning strategy around catching students after the work is submitted. It needs to design tasks where the process matters before the final product appears.
For district leaders, this means AI policy should include instructional examples. A policy document should not only define prohibited uses. It should show what responsible use looks like across learning phases.
Before learning: AI use should be limited when the purpose is to activate prior knowledge, encounter a text, attempt a problem, or form an initial question.
During learning: AI may be appropriate when students are comparing explanations, seeking feedback, identifying gaps, practicing retrieval, or testing reasoning.
After learning: AI may be useful for revision, reflection, extension, accessibility, or transfer when students can explain what changed and why.
This sequence-based approach gives educators room to innovate without abandoning learning principles. It recognizes that AI can remove unnecessary barriers, but it also insists that some struggle is educationally necessary.
The goal is not to write a policy that says “yes” or “no” to AI in every situation. That will not hold. Tools will change. Features will change. Vendor language will change. Student practices will change.
The goal is to write policy that protects the learning sequence.
If schools do not define the human work that must happen before automation, then automation will define it for us. That is the risk. Not that AI exists, but that schools will allow the tool to quietly reorganize the learning process before educators have named what must be preserved.
Sequence matters.
Policy should say so.
