Marc Lipsitch is the Berberian Professor at Stanford University, with appointments in Infectious Diseases, Biology, and the Center for International Security and Cooperation.
In this week’s conversation, Yascha Mounk and Marc Lipsitch discuss the good, the bad, and the ugly of gain-of-function research, how scientific incentives may be encouraging risky experimentation—and the recent outbreaks of Hantavirus and Ebola.
This transcript has been condensed and lightly edited for clarity.
Yascha Mounk: There are many topics I’m excited to talk to you about, and we’ll talk a little more broadly about the landscape of public health in the United States at the moment, which is interesting. But I asked you to be on the podcast because I’ve been trying to think through a subject that I find fascinating but don’t have much expertise on, which is gain-of-function research. To start off with, what is gain-of-function research?
Marc Lipsitch: “Gain of function” is a general term that means a biological experiment that gives an organism a new function—a new ability to do something, usually by a genetic change. The type of gain-of-function research that has been controversial and raised concerns among many people, including me, is what’s sometimes called enhancement of pandemic pathogens, or dangerous gain of function, or risky gain of function—there are various other names. It is a small category, but a particularly concerning one, in which the function that an organism gains is the ability to spread more easily or cause more disease in organisms we care about: people, crops, or other organisms.
The most notorious example of it was when researchers took a strain of influenza—a flu virus that had been causing sporadic disease in people and killing a large fraction of those it was infecting—and achieved the ability of that virus, through genetic and evolutionary manipulations, to transmit readily in the air. That strain was so-called bird flu, or H5N1 flu. The reason why that’s concerning is that if you do that to a strain of flu, there’s a risk that it actually becomes capable of causing a pandemic in people—a new outbreak. We’ve all experienced a pandemic recently, of coronavirus, and a flu pandemic would be equally, if not more, damaging. The idea of these studies is that the investigators are trying to understand what makes it possible for a virus like flu to become pandemic.
Mounk: At first, that just seems like a crazy thing to do. You’re taking a virus that’s somewhat dangerous and deliberately making it more dangerous by making it more easily transmitted or by making its properties worse. That sounds like a Bond villain, crazy scientist kind of thing to do. But the purported logic is that understanding these viruses better—understanding what kind of pathogens we might face—might allow us to prepare better for pandemics, to prepare vaccines. Talk me through in detail what it is that scientists are hoping to achieve with this research, which in their mind is very well-intentioned.
Lipsitch: The goal of the research, as described by the investigators, is to find ways to characterize viruses and determine which ones are dangerous, which ones might cause a pandemic if found in nature, and which ones probably aren’t capable of that. The idea is that by doing dry runs in the lab, researchers are able to compile a list of the characteristics or genetic properties of a virus that allow it to do that.
The critique from those of us who are skeptical of that is really about both the risks and the purported benefits. On the risk side, there’s good reason to be worried about the possibility that a strain of virus created that way could infect one of the lab workers and eventually transmit beyond the lab, that it could be stolen and used for deliberate mischief, or that it could enable people to figure out how to make a dangerous virus themselves. So safety, security, and the potential for misuse are all of concern.
In terms of the purported benefits, one of the arguments that I and others have made is that you learn potentially what’s true about that particular strain of flu. But there’s good reason to believe—and even experimental evidence to show—that if you do the same thing to a different strain of flu, you might not get the same result. You might even get the opposite result: it becomes less transmissible or less able to cause widespread harm. The generalizability of the information is very limited, and that has been at the heart of the debate.
Mounk: The way to think about this seems to be as a trade-off. There’s potentially information we can get out of doing this research—which is scientifically interesting in itself, but more importantly could be helpful in trying to prevent or treat a pandemic. That’s one side of the ledger. On the other side are all the risks: the risk that somebody gets infected and the virus we’ve created ends up infecting a lot of people, the risk that somebody might steal it, the risk that the knowledge we gain in the process could reach somebody trying to create a virus for purposes of bioterrorism or something else.
Let’s go through each side of this in a little more detail. You’ve started to talk about how understanding the particular ways in which the behavior of a virus might change may not generalize. What are the other purported benefits that people claim for this? What is the strongest case that defenders of this practice would make for why we absolutely need to be engaging in this work in order to gain scientifically valuable research that might actually prevent bad pandemics in the future?
Lipsitch: The case is that if we can do it in the lab, then we understand what nature will do. Nature is constantly trying things out with mutations happening and viruses spreading in animals. If we can somehow predict what will happen, then we can prevent it.
Mounk: Perhaps you can explain to laypeople why you are skeptical about whether doing experiments on one kind of virus tells us much about different kinds of viruses. But there are also a lot of other steps in this argument that you need to assume. If we figure out that the most dangerous potential viruses have a certain shape or set of characteristics, then our whole machinery of pandemic prevention needs to be able to take precautionary steps to prepare for that kind of pandemic—to pre-produce the relevant vaccines, or to plan for certain kinds of social interventions that would stop the spread.
We hope you’re enjoying the podcast! If you’re a paying subscriber, you can set up the premium feed on your favorite podcast app at writing.yaschamounk.com/listen. This will give you ad-free access to the full conversation, plus all full episodes and bonus episodes we have in the works! If you aren’t, you can set up the free, limited version of the feed—or, better still, support the podcast by becoming a subscriber today!
If you have any questions or issues setting up the full podcast feed on a third-party app, please email leonora.barclay@persuasion.community
The part where I’m quite skeptical is that part. When you look at the experience of COVID, there was certainly a pandemic that roughly speaking was in line with what we might have expected. We’d gone through H1N1 and other kinds of viruses that behaved in roughly similar ways. It’s not like we were totally surprised by the characteristics of this virus, but all of that prior knowledge doesn’t seem to have helped us respond to that pandemic in a very rational way. Even if we know that the most dangerous kinds of pandemics are going to be roughly of a certain form and shape—and perhaps we have some degree of knowledge about that which we wouldn’t have without gain-of-function research—is that realistically going to put us in a position to prevent that pandemic, or to lessen its impact, in a way that goes way beyond what we might have been able to do without that specific knowledge?
Lipsitch: I largely agree with your skepticism, though I’ll push back on one piece of it at the end. Recently, with an undergraduate student from Harvard, where I was working until this year, we published a paper looking at the claim that finding viruses in nature is an important way to create vaccines as countermeasures. What we found was that over and over again, viruses have been found in nature and little has been done about it. Then, when there is a large-scale human outbreak—usually among people in the West, and especially among wealthy people in the West, and occasionally in other settings—vaccine development begins in earnest. Or when there’s concern about a virus as a bioterrorist agent, vaccine development begins in earnest. Simply finding a virus in nature, even with signs that it could be very dangerous, has not been enough to prompt significant development of countermeasures.
Mounk: Even if gain-of-function research shows that a certain kind of pathogen could be super dangerous—that it could be a huge problem for the world—and even if that prompts calls to develop a vaccine, that development is unlikely to actually happen. So the purported benefits from the scientific practice aren’t going to arrive.
Lipsitch: That’s an important distinction—what we studied was when people find new viruses in wild animals rather than in the lab. But the broader point is that we are pretty good at mobilizing resources when there are significant numbers of people at risk, especially in the developed world, and historically we are not very good at mobilizing resources otherwise.
Mounk: Vaccine development is an incredibly expensive, complicated, and under usual circumstances, slow process, so we can only develop vaccines against a relatively limited number of diseases. There are new technologies coming online—perhaps most importantly, mRNA vaccines, but also other kinds of approaches—that have the potential, if the regulatory environment is reformed in the right ways, to speed up vaccine development hugely.
Which way does that cut in this debate? On one hand, if we are becoming much faster at developing vaccines, then we need to know less about potential pathogens in advance, because what we should do is simply wait for them to emerge and produce actual human outbreaks. Rather than needing a five-year head start to have a chance of responding with a vaccine, we can develop those vaccines very quickly—and what we should be investing in is just the capacity to adapt and roll out vaccines at scale when that happens. On the other hand, perhaps if vaccine development has become much easier, it is becoming more feasible to pre-produce defensive vaccines against a whole range of pathogens that haven’t yet emerged in real-world outbreaks. So perhaps having knowledge about what kinds of pathogens could be more dangerous is now becoming more valuable than it was in the past, because we’re going to be able to do more about it. Which way do you think our progress in rapidly producing and adapting vaccines cuts in this debate?
Lipsitch: One of the other things we found in our study of existing vaccines and virus prospecting in nature was that there are numerous whole viral families—large groups of viruses that infect and cause widespread disease in humans—for which we have no existing vaccines. So on the broad scale, if we go beyond pandemics and look at anything that infects humans and spreads, there is a whole menu of things we could be doing to develop vaccines with the knowledge we already have.
That is not to say there’s no point in getting new knowledge. The question, to me, is whether the best way to get that new knowledge is by doing very risky experiments that tell us something narrow about a particular group of viruses, or whether there are alternative approaches to studying these viruses that would tell us something about which viruses to be concerned about, but without creating viruses that are themselves dangerous.
Mounk: The claim by people who defend gain-of-function research is that we need to understand all the potential kinds of pathogens there might be in order to develop vaccines against them. What you’re saying is that we already know there are a huge number of existing pathogens that could potentially cause pandemics, and we haven’t even developed vaccines against them. So how about we start developing vaccines against these known dangers before engaging in a really risky practice to discover other kinds of candidate pathogens against which we could pre-develop vaccines in a speculative manner?
Lipsitch: That’s right. Beyond that, there are many ways to study which viruses are potentially dangerous without creating new ones. If you want to understand what makes a flu virus dangerous, we already understand many of the components of that. We know what the major determinants of transmissibility in people are in a flu virus. So we can study individual proteins of the flu virus, ask how they work, how well different variants of those proteins work, and then look for those variants in viruses that we find in nature. It is not the same experiment and it doesn’t produce exactly the same result. But if you think about getting some large proportion of the information with none of the risk, it’s a pretty good trade-off.
One of the important points in thinking about which research to do is not to frame it as should we do experiment X or not, because there are hundreds of thousands of experiments that never get done in science. Every scientist is deciding to spend their money and time on this project and not that one, on this approach and not that one. Most experiments that would be interesting never get done. It’s just a question of where you allocate your resources.
If you think about allocating the resources that would otherwise be used for a gain of function experiment to the next most useful experiment—one that would give you similar information or a similar ability to develop countermeasures—it becomes very hard to say that the marginal benefit from the gain of function study is really enough to make it worth putting the world at risk. Not everyone agrees with that view, obviously.
Mounk: We’ve given a pretty fair hearing to some of the potential benefits from gain-of-function research, and I at least am convinced that those benefits are much smaller than some of the defenders of the practice claim. Talk us through in a little more detail some of the risks of this kind of research. The claim that defenders would make is presumably that this tends to happen in labs with high levels of security, and that therefore, even though the pathogens they create may be dangerous, the risk of them getting out into the world is quite low. Starting with that claim—how confident can we be that these pathogens really would remain contained in safe research environments?
Lipsitch: Many of the studies done in the West, at least, have been done in very carefully designed, highly skilled labs. There is a lot of global variability in laboratory safety standards—and that’s setting aside the whole issue of malicious use entirely.
On the safety aspects, much of the controversy that followed the discussion of Wuhan and the origin of the COVID virus highlighted—whatever you think of that controversy—that there have been much lower levels of containment for the same experiments in China than would be expected in the United States. In particular, viruses where you really don’t know the pandemic potential are not very well regulated.
There is also the fact that gain of function studies, which tend to appear in the most prestigious journals because they’re flashy, creates an incentive for researchers around the world to do these studies in order to advance their careers and get into the top journals—which is what all scientists try to do, and appropriately so. But we really don’t want a dynamic in which the more dangerous your experiment, the more prestigious the publication. The amount of biology we learned from those gain of function studies that appeared in Science and Nature—the top journals in the field—was comparable to many other papers that appear in much less prestigious places. I would argue they appeared in those places because of how shocking what had been done was.
Mounk: Give us a little historical context on lab leaks. We can come back to the question about the origins of COVID and the research undertaken in Wuhan. My understanding is that there’s actually a long history of significant accidental lab leaks. In the 1970s, H1N1 escaped, probably from a lab in the Soviet Union. There was an outbreak of smallpox in the United Kingdom. More recently, there were escapes of SARS in a number of labs across East Asia—in Singapore, Taiwan, Beijing, and other places. There were also serious biosafety failures at the CDC in the United States involving anthrax and other pathogens.
Historically, have we been really good at avoiding these kinds of lab leaks? Or is there a pattern of lab leaks repeatedly happening—not just in one country, not just in one culture, not just in one political system? The examples just reeled off span from the United States to the Soviet Union, from North America to East Asia.
Lipsitch: The list of prominent ones you’ve given is accurate, and there have been some bacterial pathogen leaks as well—there has been a bacterial pathogen leak in China in more recent years, and one of the foot and mouth epidemics in animals in the UK resulted from a lab leak from their highest security lab.
Humans are imperfect, and almost all of these leaks are not the result of a physical system failing. Very often the issue is that a person fails to inactivate a vial of pathogen but thinks they have, or switches the inactivated vial for one that is still live. It’s just human error, and that’s the hardest kind of thing to prevent. The skill of the lab is, of course, important. But the record is, as you say, one of repeated failures—and underreporting is acknowledged even by the people who run these labs. The incentives to report an incident or accident are very low. The Netherlands, for example, has published a paper assessing the risk of underreporting and finding it to be quite high.
Fortunately, setting aside COVID and the controversy over that, most of these incidents have not gone very far—the 1977 release of H1N1 flu being the most clear exception. But they do illustrate that infections happen and transmission can occur, as it did with SARS in one of the accidental releases in the mid-2000s. It is a real risk. Because the consequences are so high, we are just not good at thinking about that risk. Biosafety is a field mostly dedicated to protecting individual lab workers and the people immediately in the area. Biosafety on the population scale is a relatively new phenomenon, as we deal with more dangerous pathogens.
Mounk: From a philosophical perspective, that seems to make a real difference. We want to protect all scientists who work in labs and make sure they are not exposed to serious danger. But at the limit, you can say that scientists agree to work in an environment with highly infectious pathogens and, even though we should obviously do what we can to protect them, they understand there is some risk involved. They can make that decision for themselves. Here, however, we are talking not about exposing individuals to a risk they have knowingly taken on, but about thousands, or tens or hundreds of thousands, perhaps millions of people dying as a downstream consequence of choices they didn’t make and had no say in. That seems very different.
When you look at this trade-off—and that’s one of the hard things to do, one of the things that political scientists and commentators often fail to do—the potential dangers are very serious and very clear, while the potential benefits are minor and questionable. So we probably should be discouraging, or perhaps banning, this practice. Why is it that this practice is continuing despite how stark this trade-off is? Why haven’t there been serious efforts to discourage or ban gain-of-function research?
Lipsitch: That’s not entirely a fair characterization. Toward the end of the Biden administration, there was a large effort led by the National Science Advisory Board on Biosecurity, or the NSABB, which became an executive order from the Biden administration in May of 2024 that laid out a framework for very careful review of this kind of research. It was comprehensive and very much applauded by people like me and colleagues who thought it was a major step forward. It was not perfect—one of the major issues is that it’s very hard for the federal government to regulate work that it does not fund. It’s not impossible and it should be done, but it’s much easier to put guardrails around work that the federal government is itself funding.
It was a very good policy. Then, under the Trump administration, some of the same people were involved in developing an executive order that came out in May of 2025 that, had it come from any other administration—Democratic or Republican—I would have given a grade of around 95%. It put quite strong restrictions on what it calls dangerous gain of function and attempted to extend some of that regulation beyond federally funded work.
The reason I add an asterisk about which administration it came from is that even at the time, it was clear that regulation and discussions of research funding were much more complicated and destructive in many ways under this administration. When the actual implementation of that executive order happened, a number of studies were stopped that were being funded by the federal government but posed no reasonable risk—no reasonable person would say these were at risk of causing a pandemic. Some of them were types of studies that have been done for decades and were simply not what any of us have been concerned about. Some that were stopped might raise legitimate concerns, but I think it would be very hard at the moment to get funding to do dangerous gain-of-function research. The pendulum has now swung in the direction of collateral damage to other kinds of research that have no plausible risk. The Washington Post reported that the administration simply wanted more things banned because they hadn’t banned enough.
Mounk: There is obviously a much broader attack on research funding from this administration that raises serious concerns. But give me a sense of the kind of risk we are currently running from gain-of-function research worldwide. You’re saying it has become much harder to get funding for that in the United States—is there still some dangerous gain-of-function research going on in labs in the United States? Has there been an international understanding to scale back this kind of research, or is it ongoing in labs in China, Russia, Europe, and other parts of the world? To what extent have we actually been able to build a consensus that we should abstain from the most dangerous forms of this research?
Lipsitch: It’s really hard to say what’s going on because there’s no central tracking and because the definitions are different in different places. There was a very good article by Tony Mills at the American Enterprise Institute, about a year or two ago, called “How the Virologists Lost the Gain-of-Function Debate,” which made the point that the leave us alone and let us do our science attitude that some virologists were taking was really self-defeating and not sustainable, particularly after people had experienced a pandemic in their own lives and seen what it really meant.
There has been some significant shift of opinion and more attention paid to the concerns about it, at least in this country. China has a relatively new biosecurity law that may represent some move in that direction, but there are a lot of things on people’s minds and this is probably not at the top of most people’s list. The sentiment is shifting to some degree towards seeing dangerous gain of function as a problem and something not to be done, but I don’t think we’ve quite reached the point of everyone agreeing to that.
Mounk: One of the elephants in the room is COVID. Early on in the pandemic, there was a pretty concerted attempt by public health officials and parts of the scientific community to rule out the hypothesis that it could have been caused by an accidental lab leak—which would likely be related to gain-of-function research going on in particular at the Wuhan Institute of Virology. Later, it seemed as though the consensus was shifting towards thinking that it had indeed been a lab leak. A number of assessments by American and European intelligence agencies strongly suggested this was the most likely cause of the pandemic. My understanding is that perhaps the consensus has shifted back somewhat, and some people are now more skeptical of the lab leak hypothesis than they were a few years ago.
What is your assessment of how likely it is that the most consequential pandemic in 100 years stemmed from an accidental lab leak? How should people who are not scientists think through this question?
Lipsitch: The best way to understand this discussion is that there is no definitive evidence either way. People who hold strong prior beliefs really have nothing more than that, because the evidence is so lacking in both directions. There is some circumstantial evidence in each direction. I am totally happy to have an honest discussion with somebody who thinks it’s 99% either way. It’s the 100% that is the problem—and opinion is roughly split between the lab leak hypothesis and zoonotic or animal origin. Certainty in either direction is just an implausible reading of the evidence. People can interpret the available signals in whatever way their priors allow, and that’s understandable. But anybody who is certain is simply dismissing all the flaws in their own evidence.
Mounk: Stepping back to a broader question: what should someone think about public health in the United States at this juncture? Going back to the pandemic, there was an obvious failure of public health. There are important things that public health did during the pandemic to keep the situation manageable, but on the whole, the Centers for Disease Control were built to prevent pandemics and coordinate a rational response to them. They had an enormous budget devoted to exactly this. In the first major pandemic in 100 years, there was all kinds of malfunction—from the difficulty of developing a reliable diagnostic test for COVID early in the pandemic, to the refusal to allow private labs to step into the breach during a crucial period, to often conflicting guidance about masks that changed rapidly from one moment to the next and then changed back.
Something seems to have gone wrong, and there is probably a serious need for reform. On the other side, the way in which the Trump administration is trying to blow up public health—changing guidance around vaccines for diseases like measles in ways that have already led to serious outbreaks of this highly infectious disease, and mistrusting all of science—seems to be sometimes aiming at the right problems but doing it in a way that is only likely to make those problems much worse.
How should somebody who sees the failings of some of our institutions, who takes the criticisms of them seriously, but who doesn’t want to throw the baby out with the bathwater, approach this debate? What kind of reform do these institutions need? Why are the reforms currently being imposed on them by the Trump administration going, by and large, in the wrong direction?
Lipsitch: As a first cut, there are significant problems with how we did public health during the pandemic, and almost everything that’s been done by the Trump administration would make it worse the next time. The biggest threat to our biosecurity right now is undermining trust in vaccines, which most Americans very strongly approve of. If you wanted to make it easier for an adversary to cause trouble in our country with biology, one of the best ways to do it is to undermine trust in the most beneficial public health innovation of the 20th century, which is vaccines. There are real problems, but shooting holes in the best parts of public health is a catastrophic way to address them.
In my darker moments, my sense is that this is simply about shrinking government down to the point where we can drown it in the bathtub, as Grover Norquist once said, and that it has nothing to do with reforming public health. I don’t think that’s the motivation of everybody involved, and I think there are some things that genuinely need to be done. The diagnosis that we made major mistakes in the pandemic is correct. Many of those early mistakes were under the first Trump administration and were partly due to denying that this was going to be a problem when all the experts were saying very loudly that it was and that we needed to prepare. We had a warning in December of 2019 that this could be a problem, and certainly by January it was a three-alarm fire. We didn’t get into gear nationally fast enough because there was too much denial going on.
Mounk: The story of who was taking it seriously and when was a little more complicated. There was a period in January and February where a portion of the Republican party was taking it quite seriously—Tom Cotton or some other senator was really warning about it. Meanwhile, Nancy Pelosi was giving press conferences in Chinatown in San Francisco saying that anybody who was concerned about this was essentially being racist against Chinese people, and that people should continue going out to restaurants. That was in part a response to some genuine craziness where Asian Americans were being discriminated against in a totally abhorrent way. But in January and February of 2020, the partisan politics of this were somewhat unclear, and large parts of the Democratic party were really downplaying it. It was around March and April that the very dangerous dynamic got set into place—where Trump wanted to keep the economy going and downplay the danger of the virus, and Democrats became more in favor of social distancing measures and closing things down.
Lipsitch: The details of what everyone said are hard to remember. There was a current of anti-Asian racism at that time, and I attended an event in Boston’s Chinatown in January or February that made some of the same points. But saying that this is not an occasion to express racism is a different claim from saying this is not a threat to public health that we need to be taking seriously. Some of us were able to make that distinction, and if certain politicians couldn’t, that’s a failing on their part.
Mounk: In general, we would all profit from politicians who are better able to make those kinds of obvious distinctions, which somehow they don’t seem to be.
Lipsitch: The CDC is an institution that can benefit from reform. I was working there part-time from 2021 until 2025 and saw a lot of great things, but also saw that the ability to explain ambiguity to the public has been lost to some degree. I remember from the 2009 flu pandemic there were really excellent examples of that from some of the top officials. There’s now a sense that conveying ambiguity is too hard and that overstating certainty is preferable. On the communication side, there is a lot to be done.
The same kind of conservatism is part of why the CDC didn’t allow other labs to do testing for COVID for a period. It’s an understandable impulse in public health—to try to discourage people from drawing their own conclusions—but not a good one. There is a lot that could change at the CDC. But undermining vaccination, which is really the one thing that doesn’t need to change except to become even more widely used, is a sign of bad faith and destructive impulses rather than reconstructive ones.
Mounk: The point about communication is really interesting. I never had the impression that there was bad faith involved in any of these failings. I don’t think that at any point important public health officials were lying or misleading the public out of commercial interest or because they were in the pocket of Pfizer or anything like that. I did often feel that they treated the public like children—that they thought, for the good of the country, we need to induce certain behaviors, like staying at home or not buying up all the masks that might otherwise be used by medical personnel when personal protective equipment was incredibly scarce in the early stages of the pandemic. The impulse was to communicate in such a way that better outcomes for everybody were more likely to come about. But there was neither a reckoning with the fact that people can smell dishonesty and that this approach was not necessarily likely to induce the behavior they were hoping for, nor a reckoning with the second-order effects—that if people feel misled about one thing, they are going to be much less likely to listen on the next thing, and trust erodes over time.
One of the things in the debate about masks and the early guidance that masks were not useful—that I think public health officials didn’t think through—was other disciplines of social science, like economics. The earlier the market gets the signal that there is going to be large and ongoing demand for masks, the faster factories producing other kinds of things could retrofit to produce more masks. Sitting in Atlanta, somebody at the CDC probably wasn’t thinking about that.
Is that a fair account of some of those communication failures? How can these institutions do better? How can they warn citizens about real dangers and exhort them to engage in the behaviors most likely to save the largest number of lives, but do so in a way that takes citizens seriously as reasoning adults who can make decisions for themselves?
Lipsitch: That’s exactly right, and I agree fully that this was not about financial corruption or anything like that. I would add to the list of concerns that it is fundamentally undemocratic—or anti-democratic—to deceive the citizenry, and that’s a problem in itself.
We actually need more research on how to effectively communicate in public health. If there is good research on this, it’s not being used, because we continue to see the failures. My hypothesis for such research would be that in any given moment, you are operating in an environment shaped by your previous actions. If you have been communicating with the public under normal circumstances in a way that suggests you always know exactly what’s going on, that you won’t revise your views, that the science is all settled—then people will be genuinely surprised when you say you don’t know. In fact, on many scientific and public health topics, there is real uncertainty. Very few things change as fast as a pandemic, to be fair. But there is uncertainty about what the best diet is, uncertainty about the timing of flu in a given year, uncertainty about a whole range of public health topics.
People deal with uncertainty in their lives all the time and don’t think that the person who acknowledges uncertainty is an idiot or doing their job badly. Even when a doctor says a treatment has a risk of helping and a risk of not helping, people can deal with that. The key is to accustom people to expect that public health, like all human endeavors, involves some science and some uncertainty—and that especially in fast-moving situations, guidance is going to change. Not because anyone was lying, but because the best understanding changes and there will be disagreement. All of that is something people could get used to, but it would be very hard to absorb in the middle of a crisis, because people don’t react well to crises. It’s a long-term project.
Mounk: The slogan that most galled me during the pandemic was “believe the science,” because I do believe in science and put great trust in it. But precisely what distinguishes science from systems of blind faith is that in science, no point of view is sacrosanct. There is genuine debate about most propositions, and when somebody comes in with a paper that challenges a long-held assumption in the field, the right way to respond is neither to dismiss it nor to blindly believe it, but to probe how strong the evidence is and then come to a conclusion about whether to revise our views. The slogan “believe the science” was so often invoked in a way that was actually against the spirit of falsifiability and rational inquiry—which is the very reason we should provisionally trust a scientific consensus in the first place.
The second thought comes from my own teaching practice. There are many insecure teachers who hate to be asked challenging questions by students because they fear embarrassment or loss of authority if they can’t answer. I love it when a student asks me a question to which I don’t have the answer, because when I can genuinely and authentically say, “I don’t know—do you want to look it up? Or I’ll look it up and come back next week”—I can feel the trust in everything else I say go up. I’ve just signaled that when I give an answer, it’s because I’m pretty sure I know it, and when I don’t know, I’ll say so. That actually increases the trust people have in a professor, at least in the classroom.
Lipsitch: Some of the smartest journalists about infectious disease have the same instinct. Helen Branswell, who is perhaps the smartest one that I know well and writes about infectious disease, once said that she doesn’t trust anyone who never says “I don’t know.” That’s exactly right.
Going back to the “believe the science” point—the phrase I heard more often was “follow the science,” and I think that’s even more problematic, in the sense that science tells you what is, not what you should do. There are consequences for what you should do that follow from what is, but the science doesn’t say you should stay home. The science says that if you don’t stay home and other people don’t stay home, there might be significant consequences for disease spread.
My wife and I and a colleague wrote an article in the New England Journal of Medicine in July of 2020 calling for primary schools to be reopened. My wife is an educational ethicist, this colleague is an infectious disease pediatrician, and I am an epidemiologist. That call was based on science—on an emerging and imperfect understanding that children were not major transmitters—and on the view that education is an important good not worth sacrificing on that scale for some benefit in reducing disease transmission, particularly given the other choices we were making at the time. People could debate that, and in fact they did. Most schools did not reopen in the fall of 2020. It’s not that the science says you must do that. The science says here are the consequences as best we understand them—and that understanding was changing rapidly.
Mounk: I went through a version of this myself. I wrote a viral article early in the pandemic encouraging the cancellation of large-scale events—this was, I think, March 6, 2020. At that point, we were still having mass concerts and mass political rallies. The debate wasn’t yet about whether there should be government restrictions on movement; it was just about whether we should be cancelling a lot of these privately held events. I got a very angry email from an old friend who had also been a boss of mine for a while, saying I wasn’t an epidemiologist and should stay out of it. My answer was that I’m a political theorist, and I was basing the factual assertions I was making on the best statements by people with expertise in virology. I wasn’t pretending to understand something about the properties of the virus that they didn’t understand. But the question of what consequences to draw from that is an ethical question and a political question, and as a citizen I should be able to speak to that—and as somebody trained in reflecting about normative issues, I actually have some professional expertise in doing so.
Later on, I was among the earlier people to say that now that a lot of people were vaccinated and we knew more about the properties of this virus, we should start reopening a bunch of social institutions. The people who liked me the first time disliked me the second time around. That is one of the hazards of writing.
There’s a lot of ethical reflection in the field of public health, and it often comes from one particular disciplinary angle. A lot of attention, for example, is given to the precautionary principle—ideas that I think would be very controversial in philosophy departments, where many moral philosophers would be far less convinced this is a black and white issue than the consensus in public health seems to imply. These ideas then guide public policy in ways that can backfire.
One example is that we might have been able to get a vaccine even earlier than we did—and it is an achievement how quickly we did get it—through human challenge trials. Those were ruled out as unethical. Yet if some people had been willing to expose themselves to a pathogen that was dangerous but unlikely to kill any one individual, we might have been able to save hundreds of thousands of lives. It is not at all obvious to me that that would have been unethical, but there seemed to be a relatively broad consensus within public health that it was.
More broadly, there was a very interesting story about somebody designing a custom-made vaccine for their dog’s tumor in Australia with the help of a number of AI systems. They said that over 50% of the work they put into this effort was filling out the enormous amounts of paperwork required to get permission to inject a novel drug into a dog that already had a likely lethal form of cancer. Do you think that in the field of public health, while in some areas like gain-of-function research we are continuing with really dangerous practices, in other areas—like human challenge trials, or the bureaucratic burdens for getting drugs to patients who are very high risk of harm in any case—the rules are too strong to actually maximize the potential benefits of medicine?
Lipsitch: It’s interesting you bring up human challenge trials, because with a moral philosopher, Nir Eyal, and another epidemiologist, Peter Smith, the three of us wrote one of the first articles advocating for human challenge trials early in COVID for exactly the reasons you describe. In retrospect, the benefit would have been pretty small, because the United States, Brazil, South Africa, and the UK—where vaccine trials turned out to be run—provided the world with a massive global public good, which was a large enough epidemic to test vaccines in. We would rather have not provided that good, but we did.
Mounk: Because in this particular case so many people were infected in the first place, there was less need for the human challenge trial. One can imagine lots of other circumstances where human challenge trials could save a great many lives.
Lipsitch: There was some movement on human challenge trials—at least one member of the U.S. Congress was quite interested, and the UK did establish a human challenge model, though somewhat later. It’s a complicated example. But the field of research ethics as a whole is very precautionary and much more concerned about avoiding harm than about doing good. Some people have begun to push back against that. Alex London wrote a book about two or three years ago that pushes back in the right direction. It has been recognized as a problem, but the level of reflection about basic principles of ethics is not that high in an applied field—it’s more about how to apply what has been taught, and what has been taught is largely non-maleficence, beneficence, and justice. Those principles are interpreted, in the standard way, to mean it is better to miss the opportunity to do something great than to cause harm. The conservatism is a real thing.
In the rest of this conversation, Yascha and Marc discuss the recent outbreaks of Hantavirus and Ebola—how concerned we should be, how the general public may respond to pandemics in future, and when it’s time to worry about the current outbreaks. This part of the conversation is reserved for paying subscribers…
Listen to this episode with a 7-day free trial
Subscribe to Yascha Mounk to listen to this post and get 7 days of free access to the full post archives.
