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 1: This post begins the series by focusing on the gap between completed work and demonstrated understanding.
The Illusion of Completion: Why Flawless Student Work Can Signal Zero Learning
A polished student product used to tell educators something important. It suggested that a student had read, thought, drafted, revised, and made choices. It was never perfect evidence of learning, but it gave teachers a reasonable window into the student’s understanding.
Generative AI has changed that assumption.
A student can now submit a clean essay, a fluent reflection, a complete lab report, or a well-organized presentation without doing the cognitive work that the artifact normally represents. The grammar may be strong. The structure may look mature. The vocabulary may be sophisticated. Yet when the student is asked to explain the argument, defend the evidence, or describe the process, the understanding may not be there.
This is not only an academic integrity problem. It is a learning problem.
Learning scientists have long warned that performance and learning are not the same thing. Students can sometimes produce correct answers without developing durable understanding. Manu Kapur calls this “unproductive success,” a condition where the learner reaches a correct-looking outcome without the productive struggle that builds transferable knowledge. In the age of generative AI, unproductive success becomes much easier to produce, harder to detect, and more tempting to normalize.
The danger is not that students will use technology. Students have always used tools. The danger is that the tool may quietly remove the exact parts of the task that were supposed to build understanding. If AI generates the thesis, organizes the outline, drafts the explanation, revises the prose, and polishes the final product, the student may only be managing the appearance of learning rather than engaging in the act of learning.
This is where pedagogical friction matters.
Pedagogical friction is the intentional preservation of the productive resistance students need in order to learn. It is not about making school harder for the sake of difficulty. It is not about nostalgia for paper worksheets or banning useful supports. It is about protecting the parts of learning that require interpretation, effort, dialogue, authorship, and accountability.
Some friction is productive. Students need to wrestle with confusing texts, compare possible explanations, revise weak claims, defend their reasoning, and sit with uncertainty long enough for understanding to develop. Other friction is exclusionary. A multilingual learner struggling to understand directions may need language support. A student with a disability may need a scaffold that removes an access barrier. The leadership challenge is not to preserve all difficulty. The challenge is to distinguish productive friction from exclusionary friction.
This distinction should change how school leaders interpret classroom evidence.
A quiet room of students producing neat digital artifacts may not be evidence of strong learning. It may be evidence of compliance. It may be evidence of efficient tool use. It may even be evidence that the most important thinking has been outsourced.
During walkthroughs, coaching conversations, and instructional reviews, leaders should ask different questions.
Can students explain how they arrived at their answer?
Can they identify what changed between their first draft and final draft?
Can they defend their evidence?
Can they explain where AI helped and where their own judgment mattered?
Can they respond when a teacher asks a follow-up question that was not part of the original prompt?
Can they show traces of thinking, not just evidence of completion?
The finished product still matters, but it cannot be the only evidence. Schools need more process-visible learning. This may include oral defenses, draft histories, peer critique, teacher conferences, handwritten planning, source annotations, revision memos, and short in-class explanations. These routines do not need to replace all digital work. They help restore the relationship between product and process.
For K–12 leaders, the larger issue is assessment design. If a task can be fully completed by AI without the student needing to think, explain, choose, revise, or defend, then the task may no longer be measuring what we think it measures. That does not mean every assignment must become AI-proof. It means assignments should become learning-centered.
The question is not, “Can AI do this?”
The better question is, “What human thinking is this task supposed to develop, and where must that thinking remain visible?”
Generative AI forces schools to confront something that was already true. Completion is not the same as learning. A polished product is not the same as understanding. When the output looks like learning, educators must look more carefully at the process that produced it.
The work ahead is not to eliminate AI from education. The work is to design learning environments where students still have to think, speak, revise, struggle, and stand behind their ideas.
That is where learning lives.
