Ascending Friction in the Classroom

In the AI Story

The pattern has been obscured by a recurrent critical template: when a new tool removes friction at one level, critics warn that the underlying skill will decline. The warning is partly correct and entirely inadequate. Computation skills did decline among students who used calculators as a substitute for understanding. But the calculator enabled mathematical education to reach problems manual computation could never have addressed. Spelling declined, and free expression expanded. Hand-revision declined, and compositional sophistication grew. Each abstraction produced real losses and larger gains, visible only when the timescale extended long enough for the ascent to manifest.

AI's ascent is distinctive because it operates at the cognitive layer. Previous abstractions removed mechanical labor; AI removes the production of the artifact itself. The student who submits an AI-generated essay has not bypassed the typing but the thinking through the argument, the discovery of what she actually believes, the encounter with the moment where the ideas resist coherence. That encounter is where divergent thinking lives; removing it extinguishes the conditions for creative development.

The ascent is not automatic. If the pedagogy does not evolve, the removed friction simply disappears—the student submits AI output as finished work, the teacher grades it as finished work, the transaction completes with no learning. The factory has produced its product; the product happens to have been manufactured by a machine, but the factory's logic cannot distinguish between human and machine production because it measures only output.

Deliberate ascent requires changing what classrooms assess. Grading the question rather than the answer is one method. Portfolio assessment that documents what the student kept, revised, discarded, and directed is another. Oral examination of the student's reasoning about AI-generated material is a third. Each method shifts assessment from production to evaluation, from output to judgment. Each requires teachers capable of evaluating higher-order thinking rather than measuring accuracy—a capacity not systematically developed by existing teacher training.

Origin

The general principle—that abstraction relocates difficulty upward—was developed in Edo Segal's The Orange Pill (2026) and elaborated across the Orange Pill Cycle. The classroom application draws on Robinson's decades of argument that educational rigor lies in depth rather than accuracy, combined with the specific conditions AI creates for redefining what assessment measures.

The concept connects to a broader tradition in learning science, including Robert and Elizabeth Bjork's work on desirable difficulties—the empirical finding that conditions producing smooth immediate performance typically produce worse long-term learning.

Key Ideas

Difficulty ascends rather than disappears. Every abstraction in the history of tools follows the same pattern: mechanical friction removed at one level relocates to a higher-order level where the tool cannot reach.

AI's ascent operates at the cognitive layer. The production of the artifact has been automated; what remains human is the judgment about whether the artifact is any good, what is missing, and what should be built differently.

Ascent requires pedagogical design. Without deliberate intervention, the removed friction simply disappears and learning collapses. The teacher's task is to construct assignments where the higher-order work becomes the visible object of assessment.

Evaluation is cognitively harder than production. For most students, recognizing what is wrong with an argument is more demanding than producing an argument—making the new pedagogy more rigorous than the old, not less.

Debates & Critiques

A central concern is whether the ascent is equitably distributed. Students who arrive at the classroom already equipped with evaluative capacity will thrive under the new assessment regime; students who arrive without it may experience the ascent as exclusion rather than growth. The pedagogical challenge is designing classrooms that develop evaluative capacity rather than presuming it, which requires investment in teacher preparation and student support that most systems have not yet made.