Practitioner Edition · ISTE+ASCD Webinar Beach Park District 3 · 2026
The Pedagogical Imperative of Friction

How to protect real learning when AI makes schoolwork easier to complete.

Presented by
Micah J. Miner, CETL, Ed.S.
For
K-12 teachers, coaches, principals & district leaders
Why productive struggle matters more than ever in the age of AI 01 / 19
A Tuesday in any classroom, 2026 02
Start with the symptom, not the theory

Before the bell rings, the work is already done.

Night before
A student asks a chatbot to draft the response before bed, then adds two sentences and a few word changes before turning it in.
During class
In high school biology, a student uses AI to generate "three thoughtful discussion questions" about cellular respiration — and asks them aloud without having worked through the reading.
3:10 p.m.
A middle school student submits a Google Classroom reflection on westward expansion: grammatically perfect, on-topic, and containing no trace of what the student actually thought.
That evening
A parent allows an AI math helper because they do not remember how to solve the problem, then wonders whether the tool explained the steps or simply carried the thinking.
Section I · Naming what we are seeing02 / 19
The right question for this webinar 03
The wrong questions are louder than the right one
×Should we ban AI in our schools?
×How do we catch students who use it?
×Which classroom AI tool is best?
→What happens to learning when students can skip the thinking and still turn in polished work?
Section I · Naming what we are seeing03 / 19
Context lens · Ong to generative AI 04
Why this is not just another tool shift

Five communication and consciousness stages.

Each stage changes what gets externalized and what learners must still do for themselves.

01

Primary orality

Ong

Knowledge lives in speech, memory, story, ritual, and communal interaction.

02

Literacy

Ong

Writing externalizes memory and makes abstract analysis, revision, and private reflection possible.

03

Secondary orality

Ong

Broadcast media restores immediacy and shared experience on top of literate systems.

04

Algorithmic orality

Miner extension

Social media keeps human content, but algorithmic systems curate visibility, reward performance, and shape participation.

05

Tertiary algorithmicity

Miner extension

Generative AI simulates dialogue and synthesis, allowing parts of reflection, debate, and composition to be outsourced.

Pedagogical friction matters because each shift can externalize more of the work that once formed human judgment.
Section I · Naming what we are seeing04 / 19
What we'll do today 05

Roadmap.

01
The classroom problem What "polished but empty" student work looks like — and why it isn't really cheating.
02
The risk: unproductive success Work that looks successful but does not build understanding.
03
A response framework: pedagogical friction The right kind of productive struggle, in four practical dimensions.
04
Leadership moves Curriculum, advocacy, and policy — what changes Monday morning.
Section I · Naming what we are seeing05 / 19
▮Part Two 06
PART 02 · NAMING THE RISK

The work looks done. The learning may not be.

When AI can produce competent academic work without the student doing the cognitive work, school risks rewarding the artifact while losing the development.
Section II · The risk06 / 19
What learning science actually says 07

Learning is not built from answers. It's built from struggle.

Decades of research — Dewey, Piaget, Vygotsky, then Kapur, the Bjorks, and Sweller — converge on the same finding.

01
Productive struggle
Schemas form when a learner has to wrestle with material that doesn't fit yet. That wrestling is the learning.
02
Desirable difficulty
Some difficulties slow performance now but make learning durable later. The Bjorks called these "desirable."
03
Productive failure
Kapur showed students who try, fail, then learn outperform those given the answer first — especially on transfer.
Section II · The risk07 / 19
Performance is not the same as learning 08

Four outcomes — only two of them are learning.

Kapur's 2×2 names the precise risk AI introduces.

Performance →
Without struggle
With struggle
Quadrant of concern

Unproductive success

Correct output. No understanding built. The artifact looks like learning — the learning never happened.

← Where AI defaults take students
Goal state

Productive success

Correct output and genuine understanding. The student wrestled with the material.

Avoid

Unproductive failure

Wrong output, no growth. Task exceeded reach without scaffolding.

Often the path to mastery

Productive failure

Wrong output, real learning. Struggle activates prior knowledge and prepares future learning.

Performance without cognitive labor
Performance through cognitive labor
Section II · The risk08 / 19
The thing to watch for in your buildings 09
Unproductive success
The artifact of learning remains visible. The process of learning becomes optional.
After Kapur (2016) · the central K-12 risk of generative AI
Section II · The risk09 / 19
Why the integrity frame keeps misfiring 10

This is not, primarily, a cheating problem.

The cheating frame

Locates the problem in a few dishonest students. Solves it with detection tools, honor codes, locked browsers.

Useful — but aimed at the wrong target.

The learning frame

Locates the problem in how the tools default to working. The path of least resistance now leads to unproductive success.

Students aren't lazy. The medium has changed.

Section II · The risk10 / 19
▮Part Three 11
PART 03 · A RESPONSE FRAMEWORK

Pedagogical friction.

The right kind of productive struggle — protected on purpose, when frictionless tools want to remove it.
Section III · Pedagogical friction11 / 19
Definition for practitioners 12
Pedagogical friction (n.)
The right kind of productive struggle — designed and protected on purpose, when the tools around us are designed to remove it.
In physics, friction is energy lost. In learning, friction is the energy through which understanding gets built.
Section III · Pedagogical friction12 / 19
The framework, four columns 13

Four kinds of friction worth protecting.

Three live in classrooms. The fourth lives in the system around them.

The Head

Thinking friction

The internal work of comprehension, synthesis, and revising your own thinking. Where understanding actually gets built.

What AI removesThe model writes the analysis. The thinking work moves out of the student's head.
The Room

Conversation friction

The dialogic work of defending claims to real, unpredictable people. Discussion, peer review, oral defense.

What AI removesChatbots reply fluently — but never disagree in the productive ways a peer does.
The World

Authorship friction

The vulnerability of being held personally accountable for one's own thinking, in front of others.

What AI removesThe link between a claim and the person making it dissolves.
The System

Policy friction

District policy, assessment design, prof. dev. — the values communicated through institutional practice.

What AI removesMost current AI policy is silent on learning science.
Section III · Pedagogical friction13 / 19
The equity guardrail 14

Not all friction is good for learning.

"Preserve struggle" without this distinction becomes "preserve inequity" wearing the language of rigor.

Preserve →

Productive friction

  • Wrestling with a difficult text
  • Defending an argument under critique
  • Composing original work in one's own voice
  • Revising in response to real feedback
  • Retrieving without notes; explaining without prompts
Reduce →

Exclusionary friction

  • Language barriers imposed on content tests
  • Inaccessible formats for students with disabilities
  • Procedural rules unrelated to learning objectives
  • Tests of stamina misread as tests of understanding
  • Norms that treat one cognitive profile as the default
Section III · Pedagogical friction14 / 19
A K-12 case the framework has to handle 15

The English Learner paradox.

A multilingual student uses an AI translator to convert a draft from their first language into English. What just happened?

Access gained

An exclusionary barrier just dropped.

The student can demonstrate content knowledge and join academic discourse without being filtered through a language barrier unrelated to the learning objective.

Voice lost

Authorship may quietly flatten.

Translation can produce grammatically correct, stylistically generic prose — removing the perspective and phrasing that make the student's writing recognizably their own.

No mechanical rule resolves this. Distinguishing productive from exclusionary friction requires situated, professional judgment — which is what this framework asks of us.

Section III · Pedagogical friction15 / 19
▮Part Four 16
PART 04 · LEADERSHIP MOVES

What changes Monday morning.

Curriculum, advocacy, and policy moves for K-12 leaders who are tired of choosing between ban and embrace.
Section IV · Leadership moves16 / 19
Three columns of moves 17

Curriculum, advocacy, policy.

AI policy should not only answer "Is AI allowed?" It should answer "What kinds of thinking must remain human?"

Curriculum

Make thinking visible.

  • Assess process, not only product
  • Sequence: struggle before AI assistance
  • Bring back oral defense, peer critique, in-class drafting
  • Require reflection on how a piece was made
  • Design tasks AI cannot complete in one prompt
Advocacy

Name struggle as a value.

  • Push back on "efficiency" as the only frame
  • Speak from technology leadership, not against it
  • Equip teachers to articulate why struggle matters
  • Defend the time productive struggle requires
  • Bring families into the conversation early
Policy

Close the learning-science gap.

  • District AI policy must reference cognition, not only risk
  • Distinguish productive from exclusionary use cases
  • Build assessment that values process
  • Fund prof. dev. on situated judgment, not tool training
  • Audit policies for equity blind spots
Section IV · Leadership moves17 / 19
For those who want the research vocabulary 18

Today's words → the research words.

If you read the qualifying paper or attend the dissertation defense, here's the same idea in scholarly language.

For Practitioners (Today)
For Scholars (the Paper)
The learning environment students now think and work inside
→
Symbolic environment
AI now generates the content students learn from and submit
→
Tertiary algorithmicity
The thinking work gets moved out of the student's head
→
Noetic displacement
Students are surrounded by fluent-sounding machine writing
→
Rhetorical saturation
Authorship and ownership become harder to see
→
Existential abstraction
The right kind of productive struggle
→
Pedagogical friction
Read the full research → minerclass.github.io / dissertationquestionsbeta / dissertation-sites
Bridge · From classroom to scholarship18 / 19
▮Closing 19
The question to take back to your building, district, or organization

What kinds of thinking must remain human — and what are we doing on Monday to protect them?

Micah J. Miner, CETL, Ed.S.
Director of Innovation & Technology · Doctoral Candidate, National Louis University
micahminer.com
Thank you · Questions welcome19 / 19