The Adaptation Gap

In the AI Story

The gap operates through multiple mechanisms simultaneously. Educational systems require years to redesign curricula, certify teachers, and graduate cohorts of students prepared for new conditions. Regulatory bodies require years to develop technical capacity, negotiate stakeholder positions, and draft frameworks that survive legal challenge. Professional communities require years to renegotiate the norms that govern recognition, advancement, and identity. Cultural narratives require even longer — the stories that make new forms of work feel meaningful typically emerge a generation after the work itself has stabilized. When the technology's development cycle is measured in months, each of these institutional timelines becomes a source of structural lag.

Juma's historical research identified the adaptation gap as the primary mechanism through which innovation transitions produce concentrated suffering. The framework knitters of the English Midlands were impoverished before the Poor Laws were reformed. The factory workers were exhausted before the eight-hour day was legislated. The scribes were scattered before the educational institutions the printing press enabled were established. The pattern is not a series of unfortunate coincidences. It is a structural feature of the relationship between technological change and institutional capacity, and the costs of the gap fall disproportionately on populations that lack the resources to navigate the transition independently.

What makes the contemporary gap distinctive is its magnitude. Juma himself recognized this in his final public statement in December 2017: "Today machines can learn to perform certain functions faster than we retrain the affected workers. This type of scenario is largely unprecedented and technologies will need to be governed differently." The governance innovation he called for has not materialized at the scale the moment requires. The educational system designed in 2024 for students graduating in 2028 was designed for a world that ceased to exist sometime in the winter of 2025. The gap is not closing. It is widening, and the widening determines the distributional character of the transition for a generation.

The gap also operates through what might be called vocabulary lag. Institutions process problems they have categories for. The standard frameworks available to institutional decision-makers — productivity measurement, labor market statistics, output quality assessment — can capture some dimensions of the AI transition. They cannot capture the dimensions Juma identified as most consequential: the degradation of tacit knowledge, the dissolution of professional community, the erosion of craft identity, the disruption of the developmental conditions under which expertise is formed. These dimensions are invisible to the standard frameworks not because they are unimportant but because they are unmeasured, and they are unmeasured because the frameworks were not designed for the phenomena.

Origin

The concept crystallized in Juma's work on agricultural innovation in Africa, where he observed that technologies transferred from developed to developing contexts routinely failed not because the technologies were flawed but because the institutional environments required to absorb them had not been built. The adaptation gap framework generalized this observation into a universal feature of innovation transitions, applicable whenever any technology moves faster than its governance environment.

Key Ideas

Structural rather than contingent. The gap is not the product of particular political failures but of the inherent difference between technological and institutional timescales.

Vocabulary precedes policy. Institutions cannot address phenomena they have no names for, and the naming process itself lags the phenomena by years.

Distributional consequences. The costs of the gap concentrate on populations that lack resources to navigate transitions independently.

Widening, not closing. The AI transition is producing the largest adaptation gap in recorded history, measured in orders of magnitude rather than percentage differences.

Juma's final prophecy. His last public observation — machines learning faster than workers can be retrained — named the condition his framework was designed to address.