Production Model versus Growth Model

In the AI Story

Franklin developed this framework through observation of how different work environments structure the relationship between output and development. The best organizations serve both models simultaneously—producing meaningful output while developing people who produce it. But the models pull in different directions. The production model asks: How much? How fast? At what cost? The growth model asks: What did the worker learn? How did her capacity develop? Is she more capable now than before? When quarterly numbers come due, when competitive landscape tightens, the production model's questions dominate because they have metrics and the growth model's questions recede because they do not.

The Berkeley study documenting AI's workplace effects captured this tension with empirical precision. By production metrics, AI tools succeeded spectacularly—workers produced more, faster, across wider scope. But researchers also documented intensification, boundary dissolution, task seepage into gaps that had served as cognitive rest. These gaps were not idle time; they were growth-model resources—moments of unstructured cognition during which the mind wanders, processes unresolved problems, makes unexpected connections. The production model classified them as waste; AI filled them with additional output; the growth model lost its substrate.

Consider the twenty-fold productivity multiplier through both lenses. Measured by the production model: unqualified triumph—same worker, same hours, twenty times the output. Measured by the growth model: a different question—what happened to the worker's understanding? The implementation work the tool replaced was not merely output but developmental experience, the patient friction-rich process through which understanding accumulates. A particular case: an engineer spending four hours daily on configuration management, tedious work she did not miss, also encountered within those hours perhaps ten minutes of unexpected difficulty forcing understanding—not the kind you read in documentation but embodied experiential knowledge depositing itself through years of encounter with the unexpected. When AI assumed four hours of mechanical work, it also consumed ten minutes of formative difficulty. The production model accounted for four hours saved; the growth model accounted for ten minutes lost. Neither captured the full picture alone.

The trade being made across AI-augmented work: developmental experience for productive output. Franklin would recognize this as characteristic of every prescriptive technology. The assembly line traded the craftsperson's holistic understanding for the factory's volume. The trade increased production and decreased developmental opportunity. The same trade now operates in cognitive work at scale and speed Franklin could not have anticipated but her framework predicts with troubling accuracy. The trade is not symmetrical—ten minutes of formative struggle cannot be restored by adding ten minutes of deliberate practice to the workday. Formative struggle is contextual, emerging from friction of real work on real problems. When friction is removed, emergence stops. A society systematically prioritizing production model over growth model in AI deployment is a society consuming cognitive capital without replenishing it.

Origin

The production-growth distinction has roots in multiple intellectual traditions—John Dewey's education as growth, Hannah Arendt's work versus labor, E.F. Schumacher's Buddhist economics—but Franklin's formulation achieved particular precision by grounding the distinction in observable workplace practice rather than philosophical abstraction. She watched how different work environments either developed or depleted the people inside them, and she identified the choice between models as structural rather than personal—embedded in evaluation criteria, incentive systems, the language used to describe success. The framework became influential in Canadian labor studies and feminist technology critique before being recovered by AI ethics researchers as a diagnostic for the current transformation.

What distinguishes Franklin's version from predecessors is her insistence that the choice is political—that prioritizing production over growth is not a natural economic law but a decision made by people with power about people without it. The decision can be contested, and the ground on which it is contested is democratic governance: do the inhabitants of the technological system have a voice in determining whether the system serves their development or merely extracts their output? This question, she argued, is the most important question any society can ask about its technologies.

Key Ideas

Two logics of work. Production model maximizes output, treating workers as means; growth model develops workers, treating process as product—the choice embedded in every incentive structure, rarely explicit but always operative.

Metrics determine which model dominates. What gets measured gets managed; production metrics (throughput, velocity, volume) are easily quantified, growth metrics (understanding, judgment, independent capability) are not—the measurable dominates the meaningful.

The formative struggle inside the tedium. Ten minutes of unexpected difficulty buried in four hours of configuration management—the production model saves four hours, the growth model loses the mechanism through which judgment renews itself.

Cognitive capital depletion. A society prioritizing production over growth in AI deployment consumes the understanding built by previous generations without creating conditions for the next generation to build their own—soil planted every season without fallow.

The choice is political, not technical. Prioritizing production over growth is not economic law but a decision made by those with power about those without—contestable through democratic governance if inhabitants have voice in the system's design.