How Virologists Lost the Gain-of-Function Debate

Shortly after the outbreak of the Covid pandemic in early 2020, rumors started to circulate that the virus that triggered it came from a laboratory in Wuhan, China. This story was quickly rejected by expert institutions and mainstream media outlets as an empirically baseless and even racist conspiracy theory — a verdict propagated through social media by fact checkers who deemed such claims to be misinformation.

The scientific community’s “overwhelming” conclusion, declared a now infamous letter in The Lancet in March 2020, was that “this coronavirus originated in wildlife,” and jumped from animals to humans. Yet as the pandemic worsened, it became clear that the alleged scientific consensus around the natural origins of Covid-19 was tenuous at best and disingenuous at worst. By the spring of 2021, the so-called “lab leak” hypothesis had gone sufficiently mainstream that it was the subject of widespread media coverage and even an official White House investigation. Rather than resolving the issue, that investigation further highlighted the divisions within the expert and intelligence communities.

The ongoing battle over “Covid origins” — a tale of deception and coverup to some, a lesson in politicization and disinformation to others — is by now familiar, if complicated and contested. While mainstream scientific opinion still holds that the virus likely originated with a natural spillover, the lab leak hypothesis is no longer dismissed out of hand and has in fact been endorsed, albeit with “low” or “moderate” confidence, by several federal agencies, including, most recently, the C.I.A. Revelations about the reckless behavior of researchers at the Wuhan Institute of Virology, combined with the uncooperative and duplicitous response of the Chinese government, have offered strong circumstantial evidence of foul play. Meanwhile, the defensive and at times egregious behavior of some within the scientific community here in the United States only exacerbated the concerns of those who suspected them of complicity.

As with so many other scientific controversies in our political life, public opinion on Covid origins has come to track — and serve as a signifier for — partisan identity. This bodes ill for dispassionate investigation, which we must have if we want to know the truth about what actually threw the world into chaos for years and killed 27 million people.

At the same time, the controversy over Covid origins thrust into the center of our culture wars a substantive debate in science policy that has been raging among experts for decades, and will continue regardless of when or whether the true origin of the virus is established. That debate turns on the risks and benefits of the very kind of research alleged to have caused the pandemic.

On the one hand are virologists, specialists in the subfield of microbiology who study viruses. Many of them have long argued that experiments in which pathogens are genetically manipulated in ways that can render them more pathogenic, virulent, or transmissible — so-called “gain-of-function” experiments — provide invaluable sources of knowledge to help us prepare for future pandemics. On the other hand are critics, including microbiologists as well as experts in biosecurity, biosafety, and public health, who have long questioned whether these experiments are worth the risk. One of their primary concerns has been that rather than helping us prepare for future pandemics, gain-of-function experiments conducted on potentially dangerous pathogens could accidentally trigger a pandemic — precisely the kind of scenario some believe transpired in Wuhan in late 2019.

In the years leading up to Covid, the debate over gain-of-function research played out in the pages of scholarly journals and inside Washington bureaucracies in a largely technical language befitting those expert institutions. The result of this long and often highly fractious dispute was a new policy framework, released in May 2024, that was designed to strike the right balance between benefits and risks. That framework was in the early stages of implementation when Donald Trump won the presidency a second time. Now, there may be no balance left to strike.

Trump — along with key advisors such as Robert F. Kennedy, Jr., his Secretary of Health and Human Services, and Jay Bhattacharya, his nominee for director of the National Institutes of Health — is poised to ban gain-of-function research via executive order. Whether such an order effectively becomes the permanent policy of the administration or gets refined by executive agencies into yet another policy framework remains to be seen. But one thing appears almost certain: the virologists who argued so vehemently for the importance of their research have suddenly found themselves completely sidelined.

How did we get here? Why did the virologists lose the debate over gain-of-function research? And what lessons can we glean from their failure? The history of this debate, and my conversations with experts on both sides of it, point to a conclusion that many in the scientific community may find hard to swallow: that the governance of gain-of-function research was never a technical problem to be solved internally by specialists themselves, however pure their motives and however valuable their expertise. Rather, it was always a political issue of public concern, requiring accountability by the scientists and moral deliberation by the country. By failing to fully grapple with this reality, the experts brought upon themselves the crisis of public doubt that was looming over them.

The term “gain of function,” little known to the public prior to Covid-19, is one of those terms that no one seems to like but is difficult to avoid, however imprecise or misleading it may be. Broadly speaking, it refers to a method of genetically manipulating organisms to confer new traits or enhance existing ones. The technique is widespread in the life sciences, and most of its uses are benign and uncontroversial. Yet the term “gain of function” itself is not strictly scientific, having first emerged in the context of policy debates over the governance of pathogen research. And it is in that context that gain of function has proved controversial.

In the spring of 2012, two earth-shattering papers appeared almost simultaneously. The first, published in the leading British science journal Nature, described an experiment by an international team of researchers supported by the U.S. National Institutes of Health (NIH). It showed that a genetically manipulated variant of H5N1 — commonly known as bird flu — could be transmitted between ferrets through respiratory droplets. A month later, the second paper appeared in the leading American journal Science, reporting the results of another NIH-supported experiment conducted by a different international team, showing that a genetically manipulated variant of H5N1 could be transmitted between ferrets via respiratory droplets or aerosols.

Together, these two experiments revealed for the first time that H5N1 — a highly deadly virus with a mortality rate of around 50 percent in humans — could become transmissible between mammals. The discovery raised the possibility that bird flu could pose a direct pandemic threat to humans. Both papers concluded by recommending pandemic preparedness, and further experiments to inform such preparedness.

But what if these experiments themselves increased the risk of a bird flu pandemic? Recall that in both experiments the researchers first had to genetically modify the virus to demonstrate the possibility of airborne transmission. The very idea that making pathogens more dangerous can help prevent future pandemics may well seem counterintuitive. Why would scientists want to take a pathogen that currently poses a relatively low risk to humans — about a thousand cases worldwide in two decades — and deliberately make it more dangerous? The conventional response is that the knowledge obtained from such gain-of-function experiments, when conducted in sufficiently high-security laboratories with appropriate safety protocols, is indispensable to pandemic preparedness and response.

Gain of function is a useful technique because it allows scientists to study characteristics of organisms they might not otherwise observe in nature. In pathogen research, its defenders maintain, the technique enables virologists to discern which pathogens pose the greatest risks and under what circumstances — which helps anticipate public health threats — and to develop tools, such as vaccines and surveillance systems, to respond to potential outbreaks.

Yet the experiments described in the two 2012 papers sparked heated debate among experts even before they were published. A foreign government, an article in Undark reported, deemed one of the paper’s contents “so risky that it could not be sent via the postal service or attached to an email.” The most prevalent concern was that by publishing the results of the experiments, the authors could be providing a blueprint for bad actors looking to build bioweapons. There were calls to restrict publication, recalling physicists’ futile effort to keep the discovery of nuclear fission secret in the lead-up to World War II.

The National Science Advisory Board for Biosecurity (NSABB), an expert panel that advises the government on biosecurity risks, initially recommended against full publication of the two papers. According to prominent microbiologist David Relman, who was on the board at the time, the unanimous decision, at least at first, was to recommend “limited publication, alerting the world to the possibility of evolved transmissibility in these viruses but with redaction of the exact genetic mutations that would enable anyone skilled in the art to synthesize these potentially dangerous viruses.” Upon further consideration, however, the NSABB reversed course and recommended publishing the two papers with some revisions, citing the research’s contribution to advancing science and protecting public health.

Supporters of that decision — including Francis Collins, then director of the NIH, and Anthony Fauci, who oversaw the National Institute for Allergy and Infectious Diseases, the institute within the NIH that funded the experiments — could appeal to longstanding norms within the scientific community such as researcher autonomy and transparency. Those who dissented, Relman among them, argued that the risks of publication were simply too high, so that traditional scientific norms must be subordinated to the wider public interest.

At issue in this controversy was thus not only a debate about the best policies for pathogen research, but a deeper philosophical disagreement about how to balance the goals of scientific progress with public safety and national security. From this point of view, the debate over the 2012 experiments can be seen as a special instance of a larger debate in American politics that goes back more than a century: What risks does the power of modern science pose to society? What is the moral responsibility of scientists to the country and the public? And does science itself equip scientists to answer these questions?

A useful place to begin this story is the development of chemical weapons in World War I. Scientists from across the United States had cooperated in researching methods for developing and protecting soldiers against poisonous chemical agents. Their work was so integral to the war effort that observers dubbed the global conflagration “the chemists’ war.”

Yet in light of the horrors of war, public sentiment turned decisively against chemical weapons, and the chemists’ work in developing them became an object of intense public scrutiny. On one side were the chemists, who argued that their inventions were not only necessary but in fact the “most humane” weapons of war because they supposedly minimized overall casualties. What’s more, they maintained, chemical warfare research could be repurposed for peacetime prosperity; it therefore deserved continued public support. The American Chemical Society went so far as to lobby successfully against the Geneva protocol banning chemical weapons.

On the opposing side was a politically heterogeneous coalition of pacifists, humanists, and religious and civic leaders who accused the chemists of shirking their moral responsibility to society. “Our chemists,” declared an editorial in The Nation, “devise more and more deadly poison gases and turn them over to governments, apparently with a clear conscience.” Critics feared that the chemists were trading on their technical expertise to monopolize public discourse to serve their own interests over the public’s, obscuring the moral questions raised by their wartime work.

This debate foreshadowed another, better-known controversy to come, over the development of nuclear weapons during World War II. Once again, scientists’ role in building weapons integral to the conduct of war — and touted as sources of innovation and abundance in peacetime — generated public backlash in the years that followed, with the arms race against the Soviet Union abroad and the effort to commercialize nuclear energy at home. It was in this context that President Dwight D. Eisenhower warned of the “unwarranted influence” of what he called the military–industrial complex and the danger “that public policy could itself become the captive of a scientific-technological elite.”

The controversies over chemical warfare and atomic energy both highlighted the “dual use” character of some forms of scientific research — its potential to be used for either benevolent or malevolent purposes. They also revealed the potential peril to scientists’ public legitimacy inherent in their collaboration with the state.

Though less well known, the U.S. government had also launched a biological weapons program in 1943, which continued on after the war. Operating largely out of the public eye until the 1960s, the program “provoked international criticism and drew public attention,” writes Jeanne Guillemin, with the “widespread use of chemicals, riot-control agents and herbicides in Vietnam.” Responding to political pressure, President Richard Nixon unilaterally ended the non-defense aspects of the program in 1969, going on to sign the Biological Weapons Convention in 1972. Yet the potential dangers of biological research would be a recurring theme in the politics of science in the coming decades.

Concerns about the technological power enabled by modern biology took on new dimensions in the 1970s with the rise of recombinant DNA, the ability to manipulate gene sequences in the laboratory. Citing the potential for biohazards that could unintentionally result from the use of such techniques — putting researchers and the wider public at risk — the National Academy of Sciences issued a letter in 1974 calling for a pause on experiments using recombinant DNA. The moratorium was voluntarily adopted by the scientific community, despite the fact that the techniques in question had “as yet no practical application” and no harm had occurred.

A conference was organized in Asilomar, California, during the pause to consider the best path forward. The result of these deliberations was a set of principles and recommendations intended to facilitate the safe pursuit of research using recombinant DNA. The storied gathering, lionized by many as an exemplary instance of ethical deliberation about the social impacts of science, has since inspired calls for caution in the development of other technologies deemed high-risk — including not only gain-of-function research but also artificial intelligence.

Yet the outcomes of the proceedings at Asilomar were ultimately overshadowed by the explosion of the life sciences industry over the course of the ensuing decades, with recombinant DNA techniques becoming utterly ubiquitous. By the 1990s, further advances in molecular biology and genetic engineering — along with the widespread dissemination of the knowledge and tools of modern genetics — raised anew the old specter that the technological power enabled by modern science could unleash dangerous threats to society, particularly if terrorists were to weaponize genetics for biological warfare.

Ian Malcolm, Michael Crichton’s iconic mathematician, vividly captured this set of concerns in the 1990 novel Jurassic Park:

Fifty years ago, everyone was gaga over the atomic bomb. That was power. No one could imagine anything more. Yet, a bare decade after the bomb, we began to have genetic power. And genetic power is far more potent than atomic power. And it will be in everyone’s hands. It will be in kits for backyard gardeners. Experiments for schoolchildren. Cheap labs for terrorists and dictators. And that will force everyone to ask the same question — What should I do with my power? — which is the very question science says it cannot answer.

After the terrorist attacks on September 11, 2001, and the anthrax attacks immediately following, the anxieties presciently highlighted by Crichton a decade before became central to the longstanding debate over so-called dual-use research. During President George W. Bush’s first term, the government initiated or expanded a number of programs to mitigate the risk of bioterrorism, from strengthening oversight and regulation of “select agents” to legislation to rapidly develop “medical countermeasures … against chemical, biological, radiological, and nuclear … threats to our national security.”

In 2004, the U.S. Department of Health and Human Services established the National Science Advisory Board for Biosecurity to provide advice on the governance of life sciences research that posed potential national security risks. It is no surprise, given this context, that the 2012 papers about the ferret experiments caused such an uproar. That same year, the U.S. government issued a policy framework at the recommendation of the NSABB for “dual use research of concern” (or DURC). It formalized a review process for government research with “certain high-consequence pathogens and toxins,” with the aim of preserving the “benefits of life sciences research while minimizing the risk of misuse of [such] knowledge.” The DURC framework emerged as one of several overlapping policies designed to balance the risks and benefits of dangerous science, including gain-of-function experiments.

The calculus of risks and benefits: this is the preferred language of the expert bureaucracy. But is it sufficient for grappling with the challenges raised by the power of modern genetics, and gain-of-function research in particular? Or does it instead offer the illusion that science can answer the very question science cannot answer: What should I do with my power?

Like the twentieth-century debates over chemical warfare and atomic energy, today’s debate over gain of function is often couched in terms of a tradeoff between scientists’ freedom and their responsibilities to society. When it comes to research that poses significant risks to society, this argument goes, the traditional norms of scientific autonomy — according to which scientists are free to choose their own lines of inquiry and methods of research and to evaluate scientific work on the basis of their own standards of evidence and excellence — conflict with scientists’ obligations to consider the social impacts of their work.

At worst, such norms can serve as ideological devices that obscure the social impacts of science and place scientists beyond the reach of ethical accountability. In this way, philosophers of science Heather Douglas and T. Y. Branch argue, the ideal of scientific autonomy comes to function as a self-serving stratagem by which scientists grant themselves a “special dispensation to be freed” from the “basic general moral responsibility” for the “foreseeable impacts of their work.” Rather than appealing to outdated norms of scientific autonomy to justify their research, scientists ought to grapple explicitly with its social implications.

This, as we saw above, is an old line of argument, dating back at least to the debate over chemical warfare after World War I, and it may seem straightforward enough in light of the power science can — and has — unleashed. Yet a common rejoinder is that while scientists may have deep and narrow expertise in given technical domains — molecular biology or organic chemistry or atomic physics — they do not have exclusive competence when it comes to translating their knowledge into practical decision-making or evaluating its social impacts. Nor does their technical expertise endow them with authority to make practical decisions that will affect society writ large.

Why should scientists, rather than members of the wider public or their elected representatives, take it upon themselves to grapple with such weighty moral questions as whether to develop weapons of mass destruction? “Such roles,” the sociologist Edward Shils once pointed out, “lie beyond the training of scientists, beyond their capacities and the legitimacy of the activities for which they claim and receive the support of society.”

The physicists who sought to retain control over atomic research in the waning days of World War II were undoubtedly acting out of concern about the social impact of their work. But it was naïve in the extreme to expect that a decision as momentous as whether and how to use an atomic weapon — or to continue developing new kinds of atomic weapons — would be left to a handful of scientists, rather than to a wartime president and his military advisors.

Or consider the famous discussion of recombinant DNA in Asilomar. Some have derided it as an unfortunate instance of precautionary reasoning, feeding into alarmism and moral panic. Others have taken almost the opposite tack: applauding the scientists’ concern for the moral dimensions of their work, but nevertheless criticizing the conference as a technocratic exercise by which experts, limiting public participation, sought to avoid burdensome government oversight by developing their own guidelines, rather than engaging in the ordinary process of political deliberation with a wider set of stakeholders.

Reflecting on the growing movement for the “responsibility of science” in the 1970s, Irving Kristol put this critique in characteristically laconic prose:

When scientists say they want to live up to their social responsibilities, what they usually mean is that they want more power than they have; it means they want to run things, to take charge. It’s always nicer to run things than to be run by them. But that’s not what moral responsibility really means.

Even if scientists ought to take up their social responsibilities, are they really equipped to do so? How can they possibly know what, if any, downstream applications of their work will prove most fruitful or destructive to society? No one in 1785 could have predicted that Claude Louis Berthollet’s discovery of the chemical composition of ammonia would lay the groundwork for Fritz Haber’s synthesis of ammonia in 1909, which would in turn enable Germany after the outbreak of World War I to circumvent the Allied blockade and continue producing fertilizers and munitions at scale. Nor could anyone in 1909 have predicted that ammonia synthesis would go on to transform global agriculture decades later in what is now referred to as the Green Revolution, enabling an expanding human population to feed itself.

Defenders of the social responsibility of science respond by conceding that some uses of scientific knowledge are not foreseeable, in which case it would be absurd to hold scientists responsible for them. But other uses of science are foreseeable — such as the development of poison gas during World War I or the development of an atomic weapon after the discovery of nuclear fission in 1938 — in which case it seems equally absurd to shield scientists from all moral responsibility. For this reason, Douglas and Branch argue, “scientists should be responsible for the foreseeable impacts of their work (rather than all impacts),” a demand that is perfectly consistent with “general moral responsibilities for all agents.”

Both sides in this debate correctly identify challenges that flow from the increasing importance of scientific expertise to the functioning of modern democratic societies. There is, on the one hand, a need to hold scientists morally responsible for the power their knowledge enables — a power that can offer both great promise and peril to society. On the other hand, and somewhat paradoxically, there is a need to avoid empowering an unaccountable scientific elite in the very effort to face up to those moral responsibilities.

Yet whichever side of this debate one takes, there is something abstract and artificial about the entire discussion, especially when it comes to pathogen research. The scientists defending the importance of gain-of-function experiments are hardly ignorant of, nor do they deny, the social impacts of their work. They are not cloistered in the ivory tower, pursuing research without consideration of its practical implications. On the contrary, they are politically active (or at least were until recently), advising policymakers and engaged in national policy arguments about why their research benefits society and thus deserves — even demands — public support, despite the risks.

To be sure, the norms of scientific autonomy remain important within the virology community, as among research scientists generally. But the debate over risky pathogen research has always operated in a highly instrumentalized context. The state supports the science in question, and scientists themselves seek that support, not mainly because gain-of-function research advances our understanding of nature, but precisely because of its perceived contributions to public health and national security, which are public goods. The disagreement therefore is not over whether but how science best serves society and what role scientific experts should play in answering that question.

Should scientists guard against public health threats by conducting potentially dangerous experiments — or should policymakers guard the public against the dangers of such research even at the risk of potentially forgoing its benefits? This disagreement — at once scientific, moral, and political — is all the more complex because of the peculiar nature of the risks that experiments on pathogens pose.

Despite important similarities to other types of risky research, the challenges posed by pathogen research are unique in several key respects. Most obviously, the controversy over gain-of-function research did not arise from scientists developing weapons of mass destruction. Whereas chemical and atomic warfare raised moral questions about the use of science to build weapons — including downstream questions about potential commercial applications — gain-of-function experiments on potentially dangerous pathogens are intended to benefit public health: primarily through the future applications of the knowledge they afford, such as developing medical countermeasures.

To be sure, concerns about the deliberate misuse of pathogen research have been part of the debate over gain-of-function experiments from the start. Like other forms of dual-use research, gain-of-function experiments pose biosecurity risks: the knowledge they generate can be used not only for good but can also be misused in ways the researchers did not intend or possibly even foresee. But over time, the policy discussion has shifted more toward a focus on biosafety risks: the risks to human health posed by mistakes that can happen in the course of carrying out these experiments.

Scientists have been aware at least since the days of Louis Pasteur and Robert Koch in the late nineteenth century that safety measures are needed to protect the scientists, physicians, nurses, assistants, and technicians who work with dangerous materials, or work in conditions that expose them to risks. Modern biosafety arguably dates to a 1955 meeting of scientists and military personnel involved in the U.S. biological warfare laboratories to develop safety protocols for dangerous research, from how to use personal protective equipment to how to sample microbes and contain chemicals. As Nicholas Evans, Marc Lipsitch, and Meira Levinson point out in a 2015 paper, these traditional biosafety concerns amount to managing occupational hazards — ensuring that the researchers and technicians working with dangerous materials are adequately protected in the laboratory.

Yet, as these authors argue, experiments with pathogens pose an additional and more complex kind of biosafety risk, because they have the potential to impact not just those who are carrying out the research but people beyond the laboratory. This fact raises ethical questions that bear some resemblance to ones raised by research involving human subjects. For example, in clinical trials for pharmaceuticals, test participants may be inadvertently harmed through the side effects of treatment, or through mistreatment. For this reason, we have well-established ethical principles, professional standards, and government policies that deal with human-subject research.

With gain-of-function research, however, those who might be negatively affected are not a delimited population of consenting test subjects; they are the public at large, all of us. This is a situation — population-level risks posed not by downstream applications but from the conduct of research itself — for which we have no similarly well-formulated and agreed-upon ethical guidelines. To make the formulation of policy even more complicated, such population-level risks from pathogen research can arise in a variety of different ways.

In the paradigmatic case, exemplified by the 2012 papers, the risk stems from the fact that the engineering of pathogens to make them more dangerous is the express goal. The risk is built in by design, even if the scientists engaging in such experiments believe that risk can be minimized or mitigated. Any harm to society would be an unintended result — for example, due to a laboratory accident — of intentionally manipulating a virus to make it more dangerous.

By contrast, in other instances, experiments involving genetic manipulation can render pathogens more dangerous to humans entirely inadvertently. In the early 2000s, for example, two Australian researchers manipulated the mousepox virus with the intention of creating a mouse contraceptive that could be used for pest control. Yet in the process, they accidentally created a more lethal strain of the virus instead. In this case, scientists achieved a result that posed a risk to society, but unintentionally.

Or consider a case where researchers are experimenting on a pathogen known ahead of time to pose a threat to humans, but they do not intend to make it more dangerous, and successfully avoid doing so. Say, for instance, they are performing a loss-of-function experiment, rendering a dangerous pathogen less transmissible. Even in such cases, though there is no gain of function at issue, a laboratory accident could in theory not only infect a researcher but also trigger a pandemic.

What these cases illustrate is the difficulty of formulating policies that capture the various sources of risk posed by pathogen research, as well as the related difficulty of defining the scope of such policies. For instance, it is not enough for policies to try to identify ahead of time which pathogens naturally pose pandemic risks to humans, since by definition gain-of-function experiments can render pathogens that may appear to pose little natural risk more dangerous. Nor is it enough to tailor policies to the technique of gain of function, since other kinds of experimentation on pathogens that pose pandemic risks to humans can be dangerous too. Nor is it enough to focus policies on researchers’ intent, since we know that unintended experimental results can pose pandemic risks no less than intended ones.

The peculiar nature of the risks posed by pathogen research has led some biosafety advocates to suggest that the field is unique. For this reason, David Relman has argued, we should not simply ask whether such research should be supported or published, as we might with other kinds of dangerous science. Instead, “we need to place greater attention on decision-making at the earliest stages of the research process,” and ask whether such research should even be undertaken in the first place.

In the summer of 2014, several high-profile incidents in U.S. government laboratories thrust the issue of biosafety into the public eye. In June, researchers working at the Centers for Disease Control and Prevention in Atlanta were accidentally exposed to anthrax due to poor adherence to lab safety protocols. In July, viral samples contaminated with H5N1 were inadvertently shipped from the CDC to a Department of Agriculture lab in Athens, Georgia. That same month, unsecured samples of the smallpox virus were discovered sitting in a cardboard box in a refrigerator on the NIH’s Bethesda, Maryland, campus.

Later that fall, responding to the political pressure resulting from these events, the government announced a temporary pause on funding of gain-of-function research on influenza, Severe Acute Respiratory Syndrome, and Middle East Respiratory Syndrome. The stated rationale was to allow time to conduct a “comprehensive assessment of gain-of-function research” with the aim of formulating a new policy framework for federal funding.

None of these biosafety incidents concerned the genetic manipulation of potentially dangerous pathogens, nor were they the first such incidents in the United States or elsewhere. But they were important to the gain-of-function debate, as Evans, Lipsitch, and Levinson recounted in their 2015 paper, because they provided “evidence of the potential for mishaps that could lead to accidental infections in even the most respected laboratories.” Indeed, one of the CDC laboratories implicated in the anthrax accident was a biosafety level 3 lab, the second most secure category. This underscored biosafety advocates’ point that the risks from experimenting on dangerous pathogens come not only from lax rules and regulations but also from human error and negligence.

In the years that followed, experts clashed over the nature and extent of the public health benefits and risks posed by gain-of-function research: whether it could be conducted safely and how, whether it was necessary or there were less-risky alternatives, and what the government’s and the scientific community’s respective roles should be in formulating and enforcing policies.

A central point of contention was how to calculate and thus balance the benefits and risks. Advocates of the research did not deny that it posed biosecurity and biosafety risks. But they tended to grant far less weight than their opponents to the likelihood of laboratory accidents triggering large-scale outbreaks. The benefits, they argued, outweighed far-fetched risks. At a 2015 workshop on the risks and benefits hosted by the National Academies, the Dutch virologist Ron Fouchier responded to Marc Lipsitch’s presentation of risk calculations by stating bluntly: “I prefer no numbers rather than ridiculous numbers that make no sense.”

In addition to their practical benefits, some proponents also emphasized the innumerable and even incalculable ways that gain-of-function experiments contribute to advancing our understanding of nature. These epistemic goods, insisted Arturo Casadevall, Don Howard, and Michael Imperiale in a 2014 paper, although almost impossible to quantify, “should be included” alongside the practical utility in the tally of benefits. “If one does that, the benefits … are potentially so great as to warrant our risking more than we otherwise might,” they concluded.

Advocates of tighter restrictions conceded that the worst-case scenarios for gain-of-function experiments, such as a lab leak triggering a global pandemic, were low-probability events — although not as low as gain-of-function proponents maintained. But the consequences, they argued, were severe enough that extra precaution was still warranted. In general, both sides agreed that gain-of-function experiments posed risks, however large and however they were quantified. The real disagreement turned on whether the risks were worth it.

Some critics questioned whether gain-of-function experiments were really as practically necessary as their proponents claimed. The utility of the knowledge generated by gain-of-function experiments depends in part on our capacity to anticipate future threats by extrapolating from existing knowledge. Yet our track record for predicting future outbreaks leaves much to be desired — something that is even more evident to us now, after Covid-19. And in many instances, there may be less risky alternative experiments — such as working with attenuated rather than wild-type viruses, or relying on in vitro studies of viral components rather than the entire virus — that can yield similar benefits.

However great their practical or epistemic benefits, advocates of tighter biosafety restrictions maintained that gain-of-function experiments ought to be classified among those types of scientific research — such as clinical trials involving human subjects or experiments with human cloning — that by their very nature raise bioethical questions. As such, decisions about whether and how to carry out such experiments call for extra scrutiny — not only stringent methods of risk management but also careful moral deliberation.

In their 2015 paper, Evans, Lipsitch, and Levinson suggested that the ethical principles for research on human subjects, which originated with the Nuremberg Code developed after World War II to prevent future atrocities like the ones committed by the Nazis, should be extended to encompass gain-of-function experiments. They highlighted the code’s requirement that research be “such as to yield fruitful results for the good of society, unprocurable by other methods.” Needless to say, defenders of gain-of-function research did not take kindly to the implied comparison to Nazi science.

The result of the government’s “deliberative process” during the pause that began in 2014 was the creation of a new set of guidelines issued by the White House Office of Science and Technology Policy in the final days of the Obama presidency in 2017, and a corresponding Department of Health and Human Services policy framework unveiled later that year, which brought an end to the pause under President Trump. The new policy was called the “Framework for Guiding Funding Decisions about Proposed Research Involving Enhanced Potential Pandemic Pathogens.”

The introduction of the term “potential pandemic pathogens,” or PPPs, was intended to lend specificity to the contexts in which experiments on pathogens merited extra scrutiny — the shorthand name of the policy is Potential Pandemic Pathogen Care and Oversight, or P3CO. Going forward, if a funding agency such as the NIH determined that a project might meet the P3CO criteria — that it could be “reasonably anticipated to create, transfer, or use PPPs resulting from the enhancement of a pathogen’s transmissibility or virulence in humans” — it would get referred to a review committee to determine if the proposal was acceptable or not, or required modifications to continue. The final decision on whether to fund the project would be left to the agency.

Although P3CO clearly represented a tightening of oversight on gain-of-function research compared to the pre-pause status quo, hardly all critics were assuaged. Biosafety and security advocates took issue with various aspects of the framework, with some arguing that it paid insufficient attention to less risky alternative experiments, and others arguing that the process was insufficiently transparent and accountable to the public. There was also a question about whether the new policies had any real teeth. For instance, it does not appear that any proposals were rejected outright after P3CO was put in place. Two NIH-funded research projects that had been in motion before the pause were permitted to resume, while a third proposal was approved (with modifications) by the review committee but ultimately had its funding redirected to alternative approaches.

This was the state of affairs when Covid-19 struck: a new and controversial oversight protocol for federal funding of risky pathogen research had been put in place alongside other frameworks for dealing with dual-use research in the life sciences and dangerous biological and chemical agents. The result was a patchwork of federal laws, rules, and recommendations — layered on top of lab safety protocols and professional norms and best practices — each with distinct but overlapping jurisdictions and implicating different domains of the vast life sciences research enterprise. Even before the pandemic was in full swing, the inadequacies of this patchwork were well known to policymakers and scientists alike.

Before and after the onset of Covid, experts were worried. Some members of the virology community feared too much discretion was left to individual researchers — for example, in deciding which biosafety laboratory level was most appropriate when conducting research on pathogens that might not obviously fit into either the P3CO or dual use research of concern (DURC) frameworks. One might hope that scientists would choose to exercise precaution. But should such risks be left up to implicit norms, particularly when there may be professional incentives to take risks or even cut corners? Instead, virologist Gustavo Palacios argued at a panel discussion I moderated in February 2024, these norms should be codified into policy so that they can be effectively enforced.

Some virologists privately told me that while they believed existing security and safety protocols were robust, at least in high-security U.S. laboratories, they worried that existing policies did not, and perhaps could not, do anything to address misaligned incentives within the research community. Prestigious journals, for example, may subtly encourage researchers to conduct “wet-lab” experiments using real biological substances, rather than “dry-lab” computer simulations or other safer alternatives. Riskier research is simply more likely to be published. This dynamic, along with a culture of publish-or-perish and a preoccupation with finding and publishing “breakthrough” discoveries, might encourage researchers to pursue risky experiments even when safer alternatives would be just as scientifically valuable.

Another central concern was that high-risk experiments might slip through. One project in particular has come under intense public scrutiny since the pandemic: a set of studies on coronaviruses led by the nonprofit EcoHealth Alliance, which sub-granted some of its research to the Wuhan Institute of Virology. This research was not even considered to fall under the scope of P3CO. Amazingly, it had not fallen under the scope of the 2014 pause, either, and so had continued on through the contentious 2014 –2018 period. An explanation for this exemption, suggested by one of the project’s lead researchers, the virologist Ralph Baric, is that because the virus had already adapted to human receptors — unlike the H5N1 flu from the 2012 papers — transmissibility experiments on that virus did not qualify as gain-of-function. (Whether researchers in Wuhan were themselves surreptitiously and recklessly engaged in gain-of-function research on the virus is another matter, Baric stated last year in testimony before Congress.)

The fact that a project this controversial not only got the green light, but didn’t even warrant extra scrutiny, suggests that P3CO was far from airtight. The scope of its regulation and oversight was far too delimited, regardless of the possible role of the EcoHealth Alliance grant in causing the Covid pandemic.

The subject of Covid origins is more than a little delicate in scientific circles. Most virologists are reluctant to discuss it, even behind closed doors. But few in the expert community seriously dispute that, whatever the status of America’s policies for pathogen research going into the pandemic, the international situation was another problem altogether. One could not assume that the norms, standards, and policies for biosafety and biosecurity developed here would be adhered to — or could be enforced — in other countries that are carrying out research on potentially dangerous pathogens. This problem persists to this day — and, unfortunately, it will not be solved by a blanket ban on federal funding for gain-of-function research.

In January 2020, around the time scientists were first sequencing the genome of SARS-CoV-2, the virus that causes Covid-19, the Department of Health and Human Services charged the National Science Advisory Board for Biosecurity with reviewing both the P3CO and the DURC frameworks in order to “provide recommendations on balancing security and public transparency.” That work was disrupted immediately by the pandemic, and the board’s focus was redirected to emergency response. By the time the HHS charge was renewed under the Biden administration almost exactly two years later, the politics of Covid had radically transformed the gain-of-function debate, pulling it out of the frying pan of expert disagreement and plunging it straight into the fire of culture war.

Congress joined the fray later that year, directing the Office of Science and Technology Policy to review federal policies on enhanced pathogens with pandemic potential. In March 2023, the NSABB finally published its recommendations, providing the basis for the latest framework, released by OSTP in May 2024 and designed to take effect in May 2025. The culmination of a year and a half of work, and nearly a decade and a half of fractious debate, the new policy established a “unified federal oversight framework for conducting and managing certain types of federally funded life sciences research on biological agents and toxins.” As of this writing, a month after President Trump took office for a second time, this framework is about to be thrown out the window, and an executive order halting gain-of-function research is widely believed to be imminent.

It may not all have been for naught. Many of the hard-fought lessons from the gain-of-function debate prior to and during the pandemic — the difficulty of defining the scope of the problem, the peculiar nature of the risks at issue, the challenge of crafting policies that address them — could prove instructive for, or come back to haunt, the effort to develop a permanent successor to past administrations’ policies. Even something as apparently simple as a blanket ban has to confront the question of what falls under its scope. But the virologists’ confident defense of the value of gain-of-function research is, to put it mildly, unlikely to get the hearing it once did in Washington, D.C.

How did this happen? To borrow from Ernest Hemingway, the virologists who lost the gain-of-function debate did so in two ways: gradually and then suddenly. From 2011, when the ferret experiments first became known to the expert community, to 2024, when the Biden administration unveiled its new policy framework, the gradual trajectory had been toward increasing scrutiny and tighter regulations of pathogen research. But then, after November, the tighter regulations turned suddenly into plans for a blanket ban.

It may be tempting to the inhabitants of mainstream scientific and media institutions simply to blame partisanship, conspiracy theories, disinformation, and know-nothing politicians who have exploited the pandemic to wage a war against science. There is no denying these forces have all played a role in the debate over gain-of-function research. But concerns about the safety and security risks posed by pathogen research can hardly be dismissed as anti-science alarmism — particularly after Covid, which revealed in tragic fashion our vulnerability to infectious disease outbreaks, regardless of their origins.

The pandemic threw the debate over biosafety into such a different light because it demonstrated what the risks of a lab leak, however we quantify its probability, could look like in the real world. As Michael Imperiale and Arturo Casadevall, two of the most prominent advocates for gain-of-function research, put it in August 2020, the “anti-GOF camp’s central argument that these experiments were too dangerous to conduct because humanity was too vulnerable to a pandemic proved correct in the sense that the world was unprepared for COVID-19.”

The failures of our pandemic response also lend credence to the argument that, given the risks posed by gain-of-function experiments, resources would be better allocated elsewhere. The weaknesses, insecurities, and points of failure in our health care system are evidently myriad. And most stem not from inadequate virology but from ordinary humdrum problems such as supply chain bottlenecks, convoluted lines of political authority, inflexible institutions, and human error and mistrust. Given that resources are limited, why not turn away from risky pathogen research and prioritize shoring up our health care infrastructure, repairing our political institutions, or funding research more directly related to the development of medical countermeasures, such as universal vaccines or antiviral drugs?

Looming over this whole debate, of course, is the unresolved question of Covid’s origin. For those who believe that the pandemic began at a laboratory in Wuhan that was conducting gain-of-function research with NIH support, the backlash against the virologists, and indeed the scientific community more broadly, is both unsurprising and well-deserved. It hardly requires explanation. Yet the overconfidence exhibited by many on both sides of the debate notwithstanding, the reality is that half a decade later we still do not know for sure the origin of Covid-19. We may never know.

But public policy cannot wait. Even if someday we were to learn definitively that the pandemic did originate in nature, the mere fact that a lab leak cannot now be ruled out dramatically reveals what is at stake in the debate over the governance of life sciences research.

Like the power unleashed by chemistry and by nuclear physics in the twentieth century, the power unleashed by modern genetics is Promethean, raising questions about whether scientists alone can be entrusted with it. To a growing number of Americans, scientists are not trustworthy stewards of a public good but members of a self-regarding clique, advancing their interests over and against those of ordinary citizens — even at the risk of inflicting disaster on society. The populist backlash to pandemic governance, followed by Donald Trump’s re-election, have forced the scientific community to confront this political reality.

Why didn’t scientists do so sooner? Prior to the pandemic, the virologists never made a forceful case to the public for why gain-of-function experiments are so vital to the public interest as to require that they be not only conducted at all but funded by taxpayer dollars. Most members of the public were not even aware of the existence of gain-of-function research, much less the reasons why anyone was doing it. This is not to say that the virologists were unconcerned with the social impact of their work. On the contrary, they were engaged in a protracted policy debate, under the aegis of governmental institutions, about the risks and benefits of their research. The expert debate over gain-of-function experiments reflected not a failure of engagement but of a particular mode of engagement.

This was a debate that played out largely among experts and within expert institutions. In this setting, the temptation naturally arises for both sides to treat tools like risk–benefit analysis as if they could resolve the debate and provide an objective foundation to policy decisions. Yet however useful and informative these tools can be, they are invariably subject to contestation and rival interpretations, as both sides have at times acknowledged. Hence the debate can quickly devolve into a disagreement over the tools of risk assessment, obscuring the fact that the issues at stake are fundamentally moral and political.

Ultimately, the question of how much risk a society should be willing to bear to advance scientific knowledge or to use that knowledge to protect against future public health threats is a question that can be illuminated by scientific experts, but not settled by them. Answering it requires not more sophisticated calculations, but richer moral deliberation. That, too, cannot guarantee optimal policy outcomes. But there is no single, optimal policy outcome in moral conflicts over what goals we want to achieve and what we are willing to risk to achieve them. Had scientists, in addition to arguing among themselves, engaged the wider public in such a political debate about ends, and not just means, they might have had the chance to establish a more enduring policy.

Of course, doing so would have exposed the virologists engaged in this debate to the political risks inherent in compromising with other stakeholders at the bargaining table, and they almost surely would not have gotten everything they wanted. But the path they chose was evidently not without political risk either. The virologists’ failure to integrate the probability of a second Trump win into their political calculus, as one longtime critic of gain-of-function research cleverly put it to me, bears a striking resemblance to their reluctance to take seriously the probability of a lab leak when assessing the risks of their own research. The gain-of-function debacle thus offers a cautionary tale not about the abdication of scientists’ social responsibilities, but rather about what happens when scientists grapple with those responsibilities in the wrong way — insularly, out of public view, and without adequate attention to the larger moral and political issues at stake.

The inevitable rejoinder to the claim that the wider public should have had a greater say in formulating policy about science is that lay citizens lack the requisite expertise. This point is true as far as it goes, but it doesn’t go very far. The modern scientific enterprise can only operate with public dispensation — all the more so when it comes to science that has the power to put the public at risk, as the Covid pandemic made all too plain. Failure to heed this lesson goes a long way toward explaining why the virologists lost the gain-of-function debate — and why so many Americans have lost their faith in scientists in general.