Should Research Prioritize the Greatest Number?

Research serves society partly by expanding knowledge and partly by addressing problems that affect people’s lives. Because time and funding are limited, researchers and institutions must make choices; it is reasonable to ask who is likely to benefit from each choice. Yet counting prospective beneficiaries cannot be the sole measure of a topic’s worth. I argue that research priorities should give substantial weight to broad public benefit, while protecting a place for urgent needs of smaller groups, promising work whose eventual value is uncertain, and researchers’ freedom to pursue questions. This balanced approach treats impact as a responsibility, not as a numerical test that every project must pass.

The case for considering scale is strongest when decision-makers compare projects that are similarly feasible and timely. If one line of inquiry could plausibly improve a shared system used by many people, while another offers a small advantage to a few, the broader project may produce more total benefit. But a head count alone can mislead: the size of a group says nothing by itself about the severity of its problem, whether it has alternatives, or how much a project could help. Thus, the number affected should be one factor in a transparent comparison, alongside the depth and likelihood of benefit and the costs of delay.

History illustrates both the value and the limits of framing research and policy around broad needs. The Social Security Act, signed in 1935, established several forms of assistance, including old-age benefits, unemployment insurance, and aid for certain groups. Its design also connected support to administrative rules and funding arrangements: old-age insurance relied on wage and payroll taxes, while states could receive federal matching funds for old-age assistance. These features show that a large public purpose does not settle how benefits should be delivered or who can access them. A priority can reach many people while still requiring careful attention to eligibility, implementation, and fairness.

The same law makes a stronger point about people whose needs could be obscured by averages. It required an opportunity for a fair hearing for applicants denied state old-age assistance. That safeguard does not establish that every decision was fair in practice; rather, it demonstrates why formal avenues to challenge a decision matter when broad programs set individual eligibility. Research priorities deserve a comparable habit of scrutiny: investigators and funders should ask whether a proposed benefit reaches people with fewer resources, and whether the evaluation overlooks a group whose need is serious but numerically small. Attention to such cases improves the quality of the public-interest judgment instead of competing with it.

A second limit on simple counting is uncertainty. Some research may not yield an immediate application, and it is difficult to know in advance which lines of inquiry will later prove useful. If funders favor only projects with easily counted, near-term beneficiaries, they risk narrowing the range of questions studied and mistaking what is readily measurable for what matters. This is not a claim that every speculative project deserves support. It is a reason to reserve a portion of research capacity for exploratory work, assessed by intellectual promise, sound methods, and the value of the question—not by a guarantee of large short-term returns that no honest forecast can provide.

Public benefit can also require attention to specific risks, not just the largest possible audience. NIST’s AI Risk Management Framework offers an example of organizing work around the practical needs of those who design, use, and assess AI systems. NIST developed the framework through collaboration between its Information Technology Laboratory AI Program and private- and public-sector participants. The framework is intended for voluntary use and to incorporate trustworthiness considerations throughout AI systems’ lifecycle. This example supports a qualified priority rule: work with potential reach across many settings can be valuable, but its usefulness depends on whether it helps people recognize and manage concrete risks. Reach alone is not a sufficient measure of success.

Moreover, a framework’s development and later adaptation show why priorities should remain open to revision. NIST released AI RMF 1.0 after a consensus-driven process that included public input; it later released a generative AI profile to help organizations identify distinctive risks and consider responses. These facts do not prove that one process is best for all research. They do show that public priorities can be informed by participation and adjusted as a field develops. A sensible funding system should therefore make its criteria visible, invite informed criticism, and periodically reconsider whether its choices still serve the people they were meant to help.

A serious objection is that emphasizing social benefit may constrain researchers’ independence or encourage funders to favor popular topics over difficult, unpopular ones. This danger is real if “benefit” becomes a demand for immediate, easily demonstrated impact. The answer is not to abandon public accountability, but to distinguish responsible priority-setting from directing every scientific question toward a predetermined outcome. Institutions can explain why some work receives support while preserving room for investigator-led inquiry and for projects serving neglected groups. Broad benefit should guide the distribution of resources, not dictate the conclusions researchers must reach.

Ultimately, researchers and funders should prioritize topics with the potential to benefit many people, but only as part of a wider judgment. They should compare the scale, seriousness, likelihood, and timing of expected benefits; protect attention to urgent minority needs; and leave room for work whose future contribution cannot yet be counted. They should also explain their choices and allow evidence and public concerns to change them. This approach recognizes that research resources are limited without pretending that a simple head count can settle every question. The strongest priority system seeks broad benefit while remaining alert to whom a majority-based measure may leave behind.

How the argument works

The essay qualifies its answer rather than choosing between public benefit and research freedom. It first establishes why scale matters, then tests that principle against two complications: urgent needs that may be numerically small and uncertain future value. The Social Security Act example illustrates that broad programs still need attention to access and procedural fairness. The NIST example shows how a public-facing framework can be developed collaboratively and adapted to emerging concerns. The final paragraphs address the risk that accountability becomes control, then state a practical standard: transparent priorities, multiple measures of benefit, and protected space for independent inquiry.