The Least Advantaged

In the AI Story

In the AI transition, the identification of the least advantaged requires precision that the current discourse rarely provides. The most visible casualties — the senior professionals whose expertise has been commoditized — command significant attention. Their losses are real. But they possess educational credentials, professional networks, accumulated savings, and the cognitive flexibility that decades of knowledge work tend to develop. By the Rawlsian standard, they are not the least advantaged.

The least advantaged are elsewhere. They are the administrative workers and junior knowledge workers whose roles constituted the entry points of professional hierarchies that are now being compressed. They are the data labelers in Kenya and the Philippines, whose labor makes the training of large language models possible and whose compensation bears no proportion to the value their work generates. They are the communities whose economic foundations are being restructured without their participation in the decisions that produce the restructuring. They are the students in educational systems that have not adapted, who bear disadvantages that compound invisibly over years.

The difference principle requires that institutional arrangements be evaluated from the perspective of these people specifically — not from the perspective of the technology executives who capture the gains, not from the perspective of the early adopters who exploit the tools most effectively, not even from the perspective of the senior professionals whose displacement, however painful, draws on reserves that the least advantaged do not possess. This evaluation is not sentimental; it is procedural. The difference principle is a test, and the test asks whether the current arrangement makes these people as well-off as they could possibly be.

Martha Nussbaum's extension of Rawls through the capabilities approach sharpened the identification of the least advantaged by asking not merely what people have but what they are able to do and to be. A person with formal access to AI tools but without the education, infrastructure, and institutional support to use them effectively possesses the capability only nominally. The structurally least advantaged in the AI transition are those whose nominal access is least likely to translate into substantive capability — and the institutions of the transition must be evaluated from their perspective above all.

Origin

Rawls introduced the concept of the least advantaged in A Theory of Justice (§15–16) as the reference class for the difference principle. He refined the concept across subsequent work, acknowledging that its specification required empirical judgment but insisting that the identification must be grounded in primary goods rather than subjective welfare.

Sen and Nussbaum's capabilities approach provided the most influential extension and critique of Rawls's specification, arguing that the metric of primary goods was insufficient to capture the full conditions of disadvantage — that two people with identical bundles of primary goods could face radically different prospects depending on their conversion capacities, and that justice required attention to what people are actually able to do rather than merely what resources they possess.

Key Ideas

Reference class, not category of sympathy. The least advantaged are identified structurally, through their position in the distribution of primary goods, not through appeals to compassion or identification with specific stories of suffering.

Evaluation from below. Every institutional arrangement must be evaluated from the perspective of the least advantaged specifically — not from the perspective of the average citizen, the median worker, or the aggregate welfare function.

Invisibility bias. The least advantaged are systematically less visible than the middle or the top because they possess fewer resources to make themselves visible; careful empirical work is required to identify them correctly.

Compounding disadvantage. Disadvantage in one primary good typically compounds across others; the educationally disadvantaged tend to accumulate disadvantages in income, opportunities, and the social bases of self-respect.

Nominal versus substantive capability. Formal access to resources or opportunities does not constitute genuine advantage if conversion capacities are absent; the least advantaged include those whose formal access is least likely to translate into substantive capability.

Debates & Critiques

The identification of the least advantaged has been contested on multiple grounds. Who counts — present citizens only, future generations, members of other societies, non-human animals? How fine-grained is the specification — socioeconomic class, intersectional position, individual circumstance? How much weight attaches to different primary goods in defining the reference class? These questions admit no uncontroversial answers, and Rawls's own responses evolved across his career. What remains constant is the moral priority: whoever the least advantaged are in a given context, the institutions governing that context must be designed to maximize their position.