The Capability Gap
The capability gap is the analytical contribution that Angus Deaton's framework, drawing on Amartya Sen's capability approach, brings to the AI distributional question. Sen distinguished between commodities — things — and capabilities — the functionings that a person can achieve. The relationship between the two is mediated by conversion factors: personal, social, and environmental conditions that determine whether a commodity translates into a genuine expansion of human functioning. Applied to AI, the distinction reveals why declining cost curves alone cannot produce democratization. A computer in the hands of a trained engineer with reliable electricity, connectivity, domain expertise, and institutional support translates into dramatic productive capability. The same computer in other hands does not. The commodity is identical. The capability it enables is radically different.
In the AI Story
The capability gap produces what Deaton's analysis identifies as the amplification paradox: AI tools amplify existing capability, which means they benefit most the populations that already possess the most capability. The engineer who directs an AI coding assistant draws on years of accumulated knowledge about systems architecture, failure modes, and the thousand small decisions that separate a prototype from a product. Without this accumulated knowledge, the AI tool produces output the user cannot evaluate, refine, or integrate into productive work. The technology is, in a precise empirical sense, anti-equalizing in its educational effects: it widens the gap between the well-educated and the poorly educated rather than narrowing it.
The capability gap is qualitatively different from the income inequalities that previous technological transitions produced. Income inequalities can, in principle, be addressed by redistribution. Capability inequalities cannot be addressed by redistribution alone, because capability depends on conversion factors — education, institutions, infrastructure, health — that cannot be redistributed the way income can. They must be built. And building them takes decades, while the AI transition is creating capability differentials in months.
The temporal mismatch is the central distributional challenge. During the intervening period — the years between the arrival of AI-driven capability differentiation and the eventual construction of the conversion factors that could narrow it — the populations that lack the conversion factors fall further behind. The falling-behind itself reduces the probability that the conversion factors will be built in time, because political attention and institutional resources tend to flow toward populations that are already succeeding rather than toward populations that are falling behind.
The capability gap operates across multiple dimensions simultaneously. Educational preparation provides the domain expertise that AI tools amplify. Infrastructure provides the physical substrate. Health status — affected by nutrition, parasitic infection, untreated mental illness, environmental stressors — determines the cognitive capacity that productive AI collaboration requires. Language availability determines whether the most capable AI systems even operate in the user's primary language. Each dimension compounds the others, and the populations disadvantaged along one dimension are typically disadvantaged along most of them.
Origin
Sen developed the capability approach in the 1980s as an alternative to welfare economics focused on income or utility. Deaton's extensive collaboration with Sen and his own empirical work on household welfare in developing economies brought the framework into mainstream development economics. The application to AI is recent but direct: the framework was designed precisely to analyze situations in which formal access fails to translate into substantive capability.
Key Ideas
Commodities are not capabilities. Access to a tool does not guarantee the capability to use it productively.
Conversion factors mediate the relationship. Education, infrastructure, health, institutional support, and economic security determine whether a commodity produces a capability.
AI amplifies existing capability. The amplification paradox means tools benefit most those who already have the most, widening rather than narrowing the underlying gap.
Capability inequality cannot be redistributed. Unlike income, capability depends on conversion factors that must be built, not transferred.
The temporal mismatch compounds the challenge. AI creates capability differentials in months; conversion factors require years or decades to construct.
Debates & Critiques
Advocates of declining-cost-curve arguments contend that the capability gap is transitional — that as AI tools become ubiquitous and as translation, interface, and infrastructure improvements accumulate, the conversion factors will matter less. Deaton's framework responds that the historical record of every previous technological transition shows the opposite pattern: the conversion factors become more determinative over time, not less, because the populations that capture early advantages use those advantages to shape the institutions that govern subsequent distribution.
Appears in the Orange Pill Cycle
Further reading
- Amartya Sen, Commodities and Capabilities (Oxford University Press, 1985).
- Amartya Sen, Development as Freedom (Anchor, 1999).
- Martha Nussbaum, Creating Capabilities: The Human Development Approach (Harvard University Press, 2011).
- Ingrid Robeyns, Wellbeing, Freedom and Social Justice: The Capability Approach Re-Examined (Open Book Publishers, 2017).