The Trump administration is centralizing and politicizing science grantmaking to an unprecedented degree. Most recently, the Office of Management and Budget proposed a rule to require political appointees to review grants and ensure the projects they fund “demonstrably advance the President’s policy priorities” and do not “facilitate” any “initiatives that … promote anti-American values.” If the new rule takes effect, the government will be able to terminate a grant if it no longer “effectuate[s] … the national interest” — in other words, if it’s no longer to the president’s liking.

The comment period on the proposed rule ended July 13. Before it can take effect, the OMB must still finalize the rule, which the agency aims to do by the end of August.

Much of the coverage of the proposed rule has focused on the ways that turning grantmaking into a political procedure is likely to dull meritocracy, weaken freedom of inquiry, and push talented scientists to leave basic research — and, for some, to leave the United States altogether.

These are all very real concerns. But there’s another alarming risk to the president’s policy: It will weaken the National Institutes of Health’s benign influence over research risks. Making science grantmaking explicitly political will hamstring the federal government’s soft power and make it less able to work with researchers to govern emerging scientific risks, like those posed by synthetic cells or specialized AI models.

The NIH has been a surprisingly central regulator of science since the 1950s, as I argue in a recent working paper. The agency has piloted regulations in domains ranging from clinical trial ethics and genetic privacy to biosecurity and genetic engineering. In almost all of these arenas, the NIH has championed self-regulatory systems in which researchers design and enforce policy alongside the government. If the NIH becomes a mere political tool, it will weaken its power to serve as a regulatory first responder, one that can work with scientists to balance the promise and peril of novel and potentially dangerous research.

Perhaps the most famous of the NIH’s interventions is the creation of the institutional review board (IRB) framework for overseeing research involving human subjects. Virtually all federally funded studies involving human subjects in the U.S. must by law receive IRB consideration. IRBs are privately operated committees that review research proposals to ensure that research comports with federally promulgated ethical principles. Contrary to the conventional narrative, the chief architect of modern clinical-research oversight was not Congress, the FDA, or a committee of expert ethicists. It was the NIH.

Although IRB consideration is today a congressionally required condition of federal funding for human-subject research, the requirement traces back to NIH actions decades in the making. The NIH invented the IRB in the 1950s as a way to navigate the ethical questions that arose from its own clinical trials involving healthy volunteers. When confronted with a series of research scandals in the 1960s, the NIH convinced its parent agency to require that NIH grantees operate local IRBs. This approach proved so influential that it was later adopted by the FDA and by Congress itself.

Or consider a second example of NIH governance: lab safety. Genetically modifying a pathogen that could cause a pandemic can be dangerous. Viruses might escape from a lab, or be intentionally released, and kill thousands or millions of people. Counterintuitively, the NIH was for decades the main regulator of genetic engineering research, and it remains a key regulator today.

Neither Congress nor the president asked the NIH to regulate genetic research. Instead, when researchers first developed gene-editing tools in the 1970s, the NIH on its own initiative worked with scientists to write biosafety rules that it turned into a condition of funding. Like human-subject rules, these rules are largely enforced by local committees. Like IRBs, these rules stuck. When Congress created mandatory oversight of the most dangerous pathogens in the 2000s, regulators adapted the NIH playbook rather than take an entirely new approach.

Both of these stories probably would have played out differently if NIH leaders and academic scientists had seen each other as adversaries, rather than as peers. Scientists would have had little reason to bring ethical questions to the NIH’s attention, or to work in good faith with the agency to design rigorous rules. And if the NIH had attempted to force researchers to adopt centralized regulations, it seems likely that the regulations would have been less expertly designed, unduly slowed invention, or collapsed in the face of congressional and university backlash. Instead, the IRB and biosafety policies piloted by the NIH have become standard across much of the world. For all their imperfections, both systems helped fill a regulatory void, and did so by drawing on researcher expertise.

Excluding external researchers from science oversight is especially ill-advised today, when we face novel, explosive ethical questions in science. Perhaps none is as significant as the biosecurity risks posed by generative AI systems. With models’ growing proficiency in biology comes their expanding potential for misuse, such as by assisting in the design or weaponization of enhanced viruses. Addressing these risks will require expertise in virology, biosecurity, AI model design, ethics, law, and other disciplines.

The NIH, in collaboration with its grantees, could be well placed to pilot new policies for governing AI models to reduce biosecurity risks, such as by studying how to regulate the movement of sensitive viral data, or methods for gatekeeping access to advanced biological AI models. But such efforts — and many other worthy grantmaker governance policies — will flounder if federal grantmakers become beholden to partisan whim rather than attached to scientific merit.

Many people are justifiably anxious that the administration’s grantmaking policies will erode U.S. scientific preeminence. We should also recognize that politicizing grantmaking threatens science grantmakers’ key role as regulators of research risks.

Doni Bloomfield is an associate professor at Fordham University School of Law and a Greenwall faculty scholar.