Russia’s Plague Mystery and the Coming Age of Bioweapons

0:06Derek Thompson: Today, a Russian plague mystery and the future of bioweapons. In January 2023, Ashish Jha, a scientist who served as Joe Biden's COVID czar, was approached after a meeting in the White House Situation Room. The situation, officials told him, was grave. American intelligence officials were seeing bioweapons research explode around the world, especially among our geopolitical adversaries, including Russia, Iran, China, and North Korea. If these states successfully built a weapon, the Biden officials asked Ja, what could we do to protect ourselves? And what should we do now to get ahead of that risk?

0:47Derek Thompson: Now, three years later, there are at least two reasons why this conversation seems unusually prescient. The first reason is that AI may pose an extreme bio-risk, as expertise in manipulating viruses and bacteria becomes the sort of thing an average person can look up on open weight models as easily as asking Kimmy or Claude to summarize a paper. The second is that recent news out of Russia finds that a mysterious death occurred at a Russian or Siberian biolab, an international incident that triggered fears of a new pandemic. According to Jha, it's not surprising that we're in a moment of elevated fear about the risk of states or non-state actors vibe coding or whipping up bioweapons.

1:30Derek Thompson: Think about the history of modern science in the last 100 years. In the late 1800s and early 1900s, chemistry took great leaps forward, including the extraordinary achievement of fixing nitrogen from the air to make fertilizer. But soon after, in the 1910s, the same scientist, Fritz Haber, who fixed nitrogen, also turned his science into a weapon as Germany deployed fatal chemical weapons on its adversaries in World War I. Then in the 1920s and 1930s, the hotspot of scientific achievement shifted from chemistry to physics, where scientists cracked open the secrets of the quantum world. One decade later, they used that science to build another weapon, the atomic bomb, in yet another world war.

2:14Derek Thompson: And now, in the 21st century, it's not chemistry or physics, but rather biology that is in the spotlight. From CRISPR to mRNA drugs to cancer therapies and GLP-1s, we are clearly and blessedly in an age of wonder in biology and the life sciences. But the modern history of science teaches us that horrors often follow wonders. Seen in this framework, if the biological breakthroughs of the last few decades become agents of war in the next few decades, that would not be an abnormal break from history. It would be the continuation of history. And of course, that is the scariest part. Today's guest is Ashish Jha.

2:57Derek Thompson: We talk about the Russian plague, the state of bioweapons research around the world, how and why AI makes this work more dangerous by making the science of manipulating biology more democratic, the optimistic argument that bioweapons might not be as dangerous as many people think, and what we should do now, not only at the level of states and societies, but the level of families. I'm Derek Thompson. This is Plain English.

3:54Derek Thompson: This episode of Plain English is presented by Comcast Business. Did you know Comcast Business detected 79.3 billion cyber threats last year? That's why they're introducing Security Edge Preferred. It's big business cybersecurity designed for small business. Get started with Comcast Business Internet for$50 a month for one year on a qualifying internet package and find out how to get one month of Security Edge Preferred on them. Plus, ask about a$500 prepaid card. Limited time offer, restriction supply, requires 300 megabits per second internet, Security Edge, Security Edge Preferred, and Wireless Connect.

4:30Derek Thompson: Call today. Ashish Jha, welcome to the show.

4:33Ashish Jha: Derek Thompson, thanks for having me here.

4:35Derek Thompson: Let's begin in Russia. A 28-year-old woman working at a laboratory in Siberia died last week of a mysterious disease. What do we know about what happened to her?

4:46Ashish Jha: Yeah, Derek, unfortunately, we know very little. What the Russian authorities have told us is now they're saying they don't think it was plague, even though there were reports that she had died of plague. They say they don't actually know why she died. And a 28-year-old woman who's otherwise healthy dying of pneumonia is pretty unusual. The fact that she worked in a lab that had a long history of biological agent manipulation, you can call it biological weapons, is concerning. So I would say we know very little about exactly what happened to her and what the Russian authorities have told us is not super reassuring.

5:29Derek Thompson: You use the word plague, and sometimes that word is used generically to mean a disease that kills lots of people. But in this case, it's used specifically to refer to a very particular disease. Can you just tell me a little bit about what that disease is and why people even thought that this was the cause of death.

5:47Ashish Jha: Yeah, yeah. So the plague is actually a specific disease. It's caused by a bacterium called the Ersinia pestis. It has been a cause of human pandemics throughout human history, right? And the Black Death. And in the Middle Ages, it was pretty devastating. It would kill a third half of the population. The good news is we have antibiotics against it. And so that's why it is no longer quite as scary, even though the word, of course, has taken on a much broader meaning. This specific lab was part of a network of labs in Russia called the Anti-Plague Network because plague occurs naturally in Siberia, as it does, by the way, in parts of the U.S.

6:26Ashish Jha: Southwest and other places. And the story, at least in the first couple of days after she died, was that she was exposed to the plague bacteria and died from that.

6:40Derek Thompson: I think one thing that confused me and confused other people is that on the one hand, Russia is saying, oh, we've got this covered. This wasn't a death from plague. Nothing to see here. But at the same time, there's reports that there was a quarantine of roughly 200 people. There were security officials that were sent to this biolab. Why such an extraordinary response if at the same time the government is saying this is essentially a nothing burger? Yeah.

7:07Ashish Jha: Well, that's the disconnect, right? So if you just think of it as, okay, a 28-year-old woman dies of pneumonia, that's tragic, but again, very unusual for a healthy 28-year-old to die of pneumonia. But then you don't generally put hundreds of people into quarantine, you don't put several hospitals into lockdown, and you don't send security officials to that city to manage it. None of that is consistent with either a simple lab accident or just a healthy young woman dying of pneumonia. All of that, of course, in the context of a lab that has a rich history of working on biological weapons. You put all that together and you understand why both U.S.

7:47Ashish Jha: government officials have been very concerned, why I have been very concerned, that there is something else that we are not being told.

7:53Derek Thompson: Yeah, in the process of my reporting for this episode, I spoke to the MIT biologist Kevin Esvelt, who estimated that, according to his analysis of historical mortality rates, a lab infection here is something like 30 to 300 times more likely than an unrelated fatal pneumonia case in a young woman like this, just as you're saying. You know, death from plague in a healthy 28-year-old woman is basically unheard of, especially if she's in a lab, which theoretically has a little bit of antibiotics in the back room. Why does it matter in this case that it was a lab infection? Why is that a particularly frightening or important possibility?

8:30Ashish Jha: Yeah. So it's not just a lab that, quote unquote, works on plague. I think you have to ask, like, what exactly does it do? And if we go back 20, 30 years to the fall of the Soviet Union and some of the lab workers and lab directors who defected around that time, they made it very clear, and our intelligence community also assesses, that these labs are working on manipulating or modifying the plague, sorry, bacteria, Yersinia, to be either antibiotic resistant or to be more deadly. we don't know if that's what was happening here. But again, if you think about a 28-year-old woman, she was almost surely vaccinated against the plague.

9:14Ashish Jha: There is a vaccine against the plague that Russia uses. I am confident they had tons and tons of antibiotics. If she actually was a lab worker there and got exposed, she would have started antibiotics immediately. And the chances that she would have died of this is extraordinarily low. So it raises the question of, was this a modified bug? Was it something completely unrelated that we don't know? What exactly is going on here? I think that's the thing that really heightens concern.

9:44Derek Thompson: Just to be clear about this lab, before we broaden the story and talk about this dawn of a new age of bioweapons research, you're saying there's a possibility that this Russian lab was engaged in both legitimate research on the plague and some kind of work on tweaking plague in order to make it more transmissible and more mortal. I mean, are we essentially just guessing here? Because at the end of the day, it's Russia, it's a big black box, and this is Siberia, which is a part of the world that is sometimes synonymous with the middle of nowhere. Or do we have a sense of what this lab is actually up to?

10:25Yeah.

10:25Ashish Jha: So I would say, obviously, we don't know for sure, but it is not just a random guess. So let me lay out the case for why this specific lab raises concerns among people who've been paying attention. First and foremost, I'll make sort of three or four points on this. One is Russia has a long history or the Soviets had a long history of bioweapons development. And this lab was part of the network of labs that were involved in bioweapons development. So there is that history here. The second piece of information is that in the last few years, our national threat assessment, our national intelligence agencies have publicly stated in reports, and there's lots of other evidence, that Russia has been substantially expanding its bioweapons programs.

11:12Ashish Jha: Now, I don't want to connect the two because I don't know. And what we don't know is, is that substantial expansion include this lab? We don't know. But given the history of this lab, given the substantial expansion by Russia in bioweapons, one has to be concerned that this lab is part of that expansion. And I think that is why this lab particularly has raised concerns and why Russia particularly has raised concerns, because basically we think there are four countries in the world that have been actively working on biological agents, dual use programs, whatever term you want to use. And Russia is clearly one of them and maybe the most prominent of them all.

11:53Derek Thompson: I want to zoom out now and consider the broader state of bioweapons in the 21st century. You've told me a remarkable story about leaving the White House Situation Room in January, 2023, and hearing something quite alarming from administration officials. Tell me what happened.

12:09Ashish Jha: Yeah. So if you go back to January 2023, that was, you know, during COVID. And it was during the big outbreak in China when China had let go of his zero COVID policy. And I was walking out of a secure briefing. And the president's Homeland Security Advisor. So the national security team at the White House reports up to two people, the national security advisor, who was Jake Sullivan, and the Homeland Security Advisor, who was Liz Sherwood Randall. Liz's job was to think about security for the homeland, for the U.S. And she and I had become very close colleagues, friends, and she basically put her hand on my arm and said, hey, can I get your help on something that's been keeping a lot of us awake at night?

12:46Ashish Jha: She has a lot of things that might keep her awake at night. And so that's not the beginning of a great conversation. Of course, you know, I said, yeah. Her next question was, remind me of your security clearance. So I reminded her of my security clearance. And that began a deep dive into learning about what at that time, of course, was information, confidential information, secure information. Now, a lot of this has gotten much more public about what we were learning about a whole new generation of biological threats that were being developed by bad actors. Or I should say state actors who who we worried about.

13:23Ashish Jha: And our intelligence agencies were particularly concerned about Russia, China, Iran, and North Korea. All of this is detailed in various national threat assessments that are public. And what we were seeing were things that were really unusual, very different than the kinds of biological weapons developments that we had seen. The one thing I will say about this, Eric, is if you think back to that time, January of 2023, this is not AI-fueled, right? I think ChatGPT had been out for a couple of months. So this was not being driven by AI. This was really much broader than that. And the question that I was asked to sort of help answer was that if these things were to be used against us, what would our capabilities be around detection?

14:12Ashish Jha: And I worked on that. But the big picture point here is that that really shifted my thinking about the global risk landscape that we had around biological threats.

14:24Derek Thompson: You told me a really interesting story about the history of modern science that connected developments in 19th century chemistry to developments in 20th century physics to developments in 21st century biology, and put these all on a string where the connective string, in fact, was weapons. Can you give me this sort of brief theory of history? Because I found it both incredibly memorable and also quite scary. Yeah.

14:56Ashish Jha: So if you think back to the three basic sciences that we all learned in high school, chemistry, physics, biology, obviously humans have been studying all three of those for millennia. But around the turn of the 20th century, late 1800s into early 1900s, we saw real breakthroughs in chemical engineering, the ability to manipulate chemistry, not just study it, but to actually manipulate it. Actually, there was a German scientist who won the Nobel Prize, Fritz Haber, who really did some of that fundamental work. Ten years later, after his fundamental work, we saw the Germans use chemical weapons in World War I.

15:36Ashish Jha: So chemical engineering comes around the turn of the century, and 10 to 15 years later, we see the use of chemical weapons. Fast forward 1920s into 1930s, fundamental work in engineering of physics, right, you know, by Niels Bohr and Max Planck and all these amazing scientists. And we learned how to manipulate physics. And about 10 to 15 years later, we saw the use of nuclear weapons in World War II. Okay, fast forward now to the third of those three natural sciences, biology. I would say it was about 10 or so years ago, maybe 15, depending on when you want to, when people like Jennifer Doudna, who invented CRISPR and Caitlin Carrico, who also won the Nobel Prize, who did the mRNA work.

16:21Ashish Jha: And these people, what these brilliant scientists did was essentially taught us how to engineer biology. That for centuries, we had learned how to study and read biology. but only in the last 10, 15 years have we actually been able to learn how to engineer biology. It would be a historical anomaly if we did not see the use of bioweapons in the upcoming years, not decades, but years. That is what history has taught us, that engineering of natural sciences inevitably leads to the use of weapons. I would hope that we can break that string, But that's, you know, in my mind, that we would have to be lucky in order to break that string.

17:03Ashish Jha: Maybe one more point on that, Derek. Sorry, just biological engineering also arrives at a time when something else has arrived, which is artificial intelligence. And I actually think that accelerates the risk, not diminishes it. The scientists that you talked about

17:21Derek Thompson: who came up with these chemistry breakthroughs in the early 20th century, they were also military inventors, right? Fritz Haber came up with the fixing of nitrogen, but also, I believe, built chemical weapons for the Germans in World War I. Leo Szilard and Oppenheimer were quantum physicists who also contributed to the creation of the atomic bomb. It is not the case, as far as I'm aware, that Jennifer Doudna and Caitlin Carrico are busy building biological weapons on purpose. Someone could steal their work and manipulate it in a way that creates a biological weapon, but it is certainly not clear that they are engaged in that activity right now.

18:08Derek Thompson: Is that the same, is that equally true for these other countries that you named, Russia, Iran, North Korea, China? Do we have an inkling, a fear, that the Jennifer Doudna's of those respective states are also working closely with the state to translate new science into new weapons.

18:30Ashish Jha: Yeah, so that's a great point, by the way. And of course, and again, I've been mentioning respect for Jennifer Downey. No one is casting a person to Jennifer Downey on this show. I won't have. Right, right. Yeah, yeah. And so the issue and the difference may be here that you actually needed that deep level of scientific expertise that only the breakthrough scientists, the people who had actually done that breakthrough work, could have brought to bear, both in the world of chemistry and physics. There are reasons to believe that you don't have to have a Jennifer Doudna-level expert. Not that places like China don't.

19:06Ashish Jha: China has amazing people, I don't know, quite at the same level. But my point is that with the proliferation of knowledge in a way that didn't exist 100 years ago or 60 years ago, the ability to copy the learnings of Jennifer Doudna or Caitlin Carrick or all of these great scientists is very different today. And there is plenty of data to suggest that great scientists in countries are indeed working with the military in those countries in these programs.

19:37Derek Thompson: What do we know about what's motivating Russia, China, North Korea, Iran to build biological weapons? And what's the most frightening thing that they could be capable of because of those motivations? Yeah.

19:52Ashish Jha: So, and I'm going to actually split those four countries into two buckets. What the threat assessment says is that Russia and North Korea clearly bioweapons, like undoubtedly. The threat assessment puts China and Iran in slightly different category, which says they have a lot of dual use programs that we know less about in terms of whether they're true biological weapons programs or not. To me, that's what the threat assessment says. I'm just going to be very honest. To me, that's a distinction without a difference. Like the truth here is that almost all of biology is dual use. And when you have people working on manipulating biology in a way that could very quickly be turned into biological weapons, you know, essentially you're there.

20:38Ashish Jha: You You don't have to have the industrial scale of production that you did with a nuclear bomb or chemical weapons because you can build biological weapons in a lab. And biology is self-replicating, so you don't need massive amounts of material. If you can infect a small number of people, it can spread very rapidly. Okay, so now coming to your question on motivation. I don't think we have, or at least I don't have a deep understanding of exactly what is motivating each of these places. For Russia, I think there has been a view that they might use these biological agents as a part of a war in Eastern Europe if there is an attempt to kind of get back some of the former Soviet countries into their fold that they wouldn't be opposed to.

21:25Ashish Jha: and it would not be something that Putin would not be willing, it would not be a red line that he wouldn't cross. So I think that's the concern. And when I talk to friends and security people in Eastern Europe, like Poland and Estonia, that's their concern around Russian bioweapons. North Korea, the concern there has been that you can literally put biological weapons on balloons and fly them over the border and release them. And obviously the South Koreans have been very concerned about this for a long time and have been paying very close attention to that issue. um iran is interesting to me because obviously iran has been in the news they have had a long interest in weapons of vast destruction they've been obviously trying to build a nuclear weapon i see biological weapons almost sort of as a cheap person's nuclear weapon um in the sense that you don't need the infrastructure you don't need the set of resources we know that um iran has had a specific lab that was working on quote-unquote dual-use programs.

22:26Ashish Jha: That lab, run by the IRGC, was bombed during our bombing back in the spring. I don't know, and I don't know if our intelligence officials know, I don't know what has happened to that lab. But that concerns me, the fact that we may not know. And China is, to me, the most interesting country because obviously, and I have said this many times in the public, I do not think that some morning Xi Jinping is going to wake up and decide to use a bioweapon. I just think that's extraordinarily unlikely. I'm not a China expert. And so what motivates their building of this program is unclear to me outside of it could be a part of a, you know, kind of a multimodal attack, maybe around Taiwan or something else.

23:12Ashish Jha: But I have to say that my understanding of their motivations is more limited.

23:17Derek Thompson: I have two follow-ups to that last comment. The first question is, does the U.S. have what a Chinese national security advisor or the Chinese equivalent of Ashish Jha would call a dual-use program?

23:35Ashish Jha: So the official, and I will tell you what I know about this, the official, you know, stance of the U.S. government is we don't have an offensive biological program. I don't believe we have a dual-use program in any substantive way. We have lots of defensive biological programs that one could misuse for offensive purposes. But the stance of the U.S. has been that we don't need one. And that's because we have so many other military capabilities that if somebody were to attack us, there is very little for us to be gained by using a biological weapon, we could, you know, we could strike back using lots of other capabilities.

24:17Derek Thompson: I'm not a biologist. I certainly don't work in an advanced biolab, but I can imagine an American national security person saying, we needed to see COVID to build the COVID vaccine. We need to understand super COVID in order to build the super COVID vaccine. We need to understand super plague, some bio-enhanced plague, in order to build the vaccine against that. And so he said something interesting, which is that we don't have an offensive program. But I can imagine a world in which a defensive program designed exclusively with the intent to build bioweapons, keep them in the lab, and then build vaccines that can protect people from those bioweapons would nonetheless be what an outsider would define as a dual-use program.

25:11Derek Thompson: Does the U.S. have that? Yeah.

25:13Ashish Jha: Well, look, I mean, we have BSL-3 and BSL-4 labs. These are the highest security labs. They work with dangerous pathogens. What are they doing with dangerous pathogens? Exactly what you described, which is they are working with dangerous pathogens to build medical countermeasures. I think of it as a little differently than when I think about the dual-use program in Iran, for instance, where based on, again, what our intelligence agencies have said publicly, they were doing the kind of research that would lead very easily to an offensive program. I would say that is not the kind of research we are doing in these labs.

25:53Ashish Jha: We really are working with dangerous pathogens. And again, under these very sort of tight security controls to understand them better and to build medical countermeasures better.

26:03Derek Thompson: So you mentioned BSL-4 and BSL-4 is biosecurity level four, which is the security level necessary to work with the most dangerous pathogens for the purpose, ideally, of developing, as you called them, medical countermeasures. So either medicines or vaccines. Yeah, like vaccines and treatments. Have there been efforts to ban this kind of work internationally, similar to, let's say, nuclear arms measures between the U.S. and the Soviet Union, or something like the Montreal Protocol, which bans dangerous chemicals creating holes in the ozone layer at an international level based on a collective understanding of the science?

26:49Derek Thompson: Do we trust our counterparties enough to sign a treaty that bans this kind of work, even if we can't perfectly verify that they're not working on their own equivalent of BSL-4 Labs 2?

27:05Ashish Jha: Yeah, great question. Actually, I'm going to answer that question in two ways. First, there's something called the Bioweapons Convention, Biological Weapons Convention, BWC. That's been around for 70, 80 years. And it basically, all these countries have signed on saying we're not going to build biological weapons. There's no verification mechanism.

27:28Ashish Jha: And my view is it's great to have a convention for which there's no verification mechanism. And basically, you really can't rely on it at all. I don't think that it's useful. I also don't think there's an easy way to do verification for it. But that has existed. And, you know, obviously, people believe in that convention. But I don't think we believe the U.S. government has said on multiple occasions that we believe countries have violated that. But there are no consequences. So having just a convention, in my view, without verification, without any teeth, without the ability to send inspectors in just doesn't do much good.

28:05Ashish Jha: You asked about labs. So let's actually talk about those labs. No one keeps track of BSL-3 and BSL-4 labs. These are the two most, right, BSL-3 and level 4, the two most interests. Nobody keeps track of them. My best assessment tracking this over the last few years is that the number of BSL-3 and BSL-4 labs in the world have doubled in the last five, seven years. There's just a lot more of them. And they're not all in the U.S. or even in China. They are all over the world. and no one tracks them. There are supposed to be, there are these published safety protocols that you're supposed to have.

28:41Ashish Jha: No one verifies whether people are using those safety protocols or are adhering to them. And one of the things that I have actually called for, and I said this when I was in the Biden administration and I've been much more vocal in the last couple of years, is a great place for us to work with China. We did this with the Soviets, right? Where we worked with them on nuclear safety. And I actually think we should put together a team. We should come up with some rules And basically, we should create a global mechanism for inspection that actually goes into these labs and looks for safety protocols and looks at safety as actually being adhered to.

29:13Ashish Jha: This would be a fantastic way for us to work with China, maybe even Russia. But we need that and we don't have that. Most of these labs are doing more or less whatever they think without any oversight. One thing I don't understand, and I'm happy to not understand it, is how in a world with more and more BSL-4 labs in various countries that are not exactly known for their operational

29:41Derek Thompson: intelligence or their operational excellence. Like, I don't know a whole lot about Russia. I'm just guessing that they don't have the same operational excellence at their BSL-4 labs as the U.S. has. It's a poorer country. It's a tyranny. It's just a very different place. Sorry to all the Russian listeners here. It's kind of amazing to me that there hasn't been more catastrophic accidents. I mean, if we're talking about a fundamental unit of danger that is not like enriched uranium, it seems very difficult to enrich uranium to the scale necessary to build an atomic bomb. It doesn't seem that difficult to have a quantity of plague and then have someone say, I want to start another lab to study plague and develop offensive and defensive measures around this plague, and then to just distribute some of that plague to the other lab in a vial.

30:39Derek Thompson: are you sometimes happily surprised by how few accidents there have been so far as i'm knocking on wood to not jinx us yeah given that we've seen an acceleration of these kind of labs throughout countries that we i would just think would make more mistakes than like the most

30:59Ashish Jha: conscientious group of scientists yes um so here's i first of all i have been pleasantly surprised And here's my explanation for why. First of all, I suspect there have been a lot more accidents than we're aware of. These places don't report the accidents. And maybe a few people die. Maybe a bunch of people get sick. And as what just happened in Siberia, but maybe it's in a place where they're able to kind of clamp down on it a lot more. Nothing has gotten out. Obviously, there are lots of people who think there was an exception to that from the Wuhan Institute of Virology back in 2019. um so and again and i think you know i believe uh and i have said this now publicly to a lot of people getting very mad at me that i think more likely than not it was a lab leak out of wuhan but look the truth is that these things happen with some frequency and we have been very very lucky and and i think there have been two things that have kind of driven that and i think they're both changing.

32:00Ashish Jha: I mean, one is engineering pathogens in ways that make them more transmissible is very hard. That make them more deadly is not quite as hard, but still pretty difficult. And the level of expertise one needed to have was quite substantial. You needed large teams of PhDs with great resources doing it. And that was not widely available. So yes, you had BSL-4 labs. I bet there have been accidents. I bet lab workers have died. I bet you and I just don't know about it. As I said in that last phrase, I alluded to the idea that I think that is now changing because of artificial intelligence. And so I think the idea that what we haven't seen in the last five, seven years, you know, sort of the past performance doesn't predict future performance.

32:50Ashish Jha: I think that has never been more true than in this issue.

32:54Derek Thompson: Right. What I hear you saying is that expertise has been the bottleneck to stop a disaster from happening. And therefore, it would be quite dangerous if that bottleneck was cleared by, say, some technology that democratized extraordinary expertise in the realm of biology. Dot, dot, dot. We're going to get to AI in 30 seconds. At the risk of derailing this conversation, which I do not want to do, let's just limit this answer to two minutes. Why do you think COVID was more likely a lab leak than a wet market event?

33:28Ashish Jha: Yeah, look, I've debated this now with like Angie Rasmussen and others. My view on this is, I actually initially did think it was much more likely to have been a natural spillover because most outbreaks in the history have been natural spillovers. I think there's a bunch of information that points to the lab. The fact that, first of all, this lab was working on very, very similar viruses. The fact that a lot of the information about exactly what experiments were being done, what labs, what viruses were there, all of that was taken down in late 2019. The fact that every attempt that WHO and others made to try to get into the lab to speak to workers were largely blocked or obfuscated in ways that made it very difficult for anybody to assess.

34:18Ashish Jha: And then there's just a lot of intelligence out there about the outbreak that, for me, doesn't mean I am sure it was a lab leak. One of the things I say is anybody who tells you they're sure is selling you something. But I went from kind of 80-20, it was likely a natural, to 80-20, it was likely a lab leak. I don't know for sure, but I've just seen so much evidence that points in that direction, that that is where I have landed. But I could easily still be proven wrong if we get more definitive information. I just worry that I'm not sure we will. And the last point I'll make on this is this has also gotten so incredibly, insanely partisanized that it's just very hard to have a conversation about this without everybody suggesting that you have a partisan motivation for saying, you know, you're on one side or the other.

35:07Derek Thompson: I think I basically agree with you. I think the balance of circumstantial evidence, in my mind, points to a lab leak. And if presented with some slam dunk evidence that it was a wet market spillover, I would say, oh, OK, I was wrong. Yeah, I think this might be one of those events that that we never truly get to the bottom of, which is itself, I think, quite unfortunate, quite scary because COVID was an unbelievably significant event, I think, in the history of the 2020s. I mean, not just in terms of the millions that it killed, but I think if you look at inflation, levels of socialization and self-isolation, changes to the economy, we are still living in many ways in a COVID age.

35:47Derek Thompson: And it's quite remarkable to me that the Big Bang moment is very much like the OG Big Bang moment. Like we know that it happened, but we actually we still don't know why it happened or what caused it. And that's, you know, as an explainer journalist, that's certainly a matter of deep frustration. Turning the page, or at least driving the car back onto the main highway here, I want to talk about how AI changes the picture of what you're describing. You argue, or I hear you arguing, that AI is destroying the barriers, or a critical barrier, that has made biological weapons so difficult to tweak in a way that makes them more transmissible or more deadly.

36:26Derek Thompson: As specifically as you can be here, what exactly does AI give someone trying to manipulate a virus or bacteria that wasn't possible five years ago? Yeah, good question.

36:44Ashish Jha: So I see this in several ways, several types of knowledge that AI is now able to give you. So, like, people still ask the question, like, can a kid in a, you know, in a basement build a bioweapon? Like, no. Like, I think that's just, I don't think so. I think it's very, very hard. But the level of expertise one needed to essentially work with and manipulate viruses was very, very substantial. Like, you didn't just need a PhD. I mean, you had to be a very well-trained virologist. now you i don't think you need quite that depth of expertise if you understand a good amount about biology let's say you have a master's maybe you have a phd um ai will give you a lot of the capabilities and give you a lot of the information you need to close that knowledge gap there's a second part that is also important which is you needed teams of people because to build biological weapons it's not like you need one kind of expert you need multiple types of experts And it was hard to pull all those different types of experts together.

37:49Ashish Jha: Again, I think what AI is allowing is allowing people who are pretty smart, who are well-trained in biology, to learn about a whole bunch of things that are adjacent to their field. This is something that I think often, and this came up a lot in COVID, right? Like, everybody was a COVID expert. Actually, there was almost no one who was a COVID expert. There were virologists who really understood the virus. There were immunologists who really understood. And so the deeper you go into expertise, you start realizing that you actually need to understand things at a much deeper level. Like one of the things I try to do, because I was not a virologist or an immunologist, is I try to talk to the best in the country and synthesize that information.

38:22Ashish Jha: And that was what I was trying to bring to the table. Um, that is what AI I think increasingly allows people to do is synthesize expertise, fill in gaps in expertise. You don't need the same number of people. You don't need the person leading it to be an Oppenheimer with deep, deep, deep physics knowledge or deep biological knowledge. and here's the other part because it has democratized us and the other part that has happened is the ability to to manipulate biology biological engineering has just become really ubiquitous if you remember the story i started with in the white house was in 2023 before ai engineering biological engineering tools are becoming pretty commonplace we have autonomous labs that can, you just upload instructions on experiments and it'll run them.

39:12Ashish Jha: All of that means lots more people are trying. Like, I don't think lots of people are trying to build atomic bombs. Like, I don't know. I'm not an expert on that, but I suspect it's not a lot of people because it's really hard. People perceive, and I think to some extent rightly, that AI has made biological weapons much easier. The problem with that is now lots and lots of people, state and non-state actors, are trying. And I think it's still difficult for any one person. But when I speak to people who think about this all the time as scientists, and when I ask them, what are the chances that no one will be successful in the next 6, 12, 18 months, pretty much everyone agrees the chances that no one will be successful is close to zero.

39:56Derek Thompson: There are lots of similarities between this conversation that we're having about bio-risk and the conversation that I had with Alex Stamos about cyber risk. Because one of the points that he emphasized over and over again is that one way that open-weight models in particular could be particularly dangerous for cyber hacking is that cyber hacking used to require enormous teams of software developers. And some people in those teams just, you know, they were criminals. They just weren't very scrupulous. They weren't very conscientious. Maybe they would just like, you know, talk to some CIA informant because they didn't know how to shut their mouth.

40:32Derek Thompson: And then suddenly, boom, the cyber hacking network has been infiltrated and shut down. He said with these new tools, though, it's creating more one man bands or one woman bands where it previously required the team of 20. But now someone can do it on their own or with a much smaller team, which reduces the risk that law enforcement stops them. turning this back to you though if expertise is becoming cheap and democratized what's still genuinely difficult in the construction of a bioweapon like i don't want to turn this into like how to make a bioweapon 101 here but you know i i presume that's not how this conversation is going to be going to go or how it's going to be used but i still want to understand for the purpose of blocking efforts, like what is genuinely hard to attain to come into ownership of if you are the sort of madman trying to build a dangerous bioweapon?

41:34Ashish Jha: Yeah, that's a great question. And obviously, by the way, again, I'm not a scientist, a biological scientist who kind of works at this at the frontier. The conversations I have, and this is a place where scientists really disagree, and I actually worried that there is a lack of imaginative thinking on this. A lot of virologists are like, oh, this is too hard. It's really, and it is very hard, right? Like manipulating viruses in a way that like, and I go back to a line that a friend of mine, Tony Fauci, often said to me when we would talk about these topics, he'd say, nature is, you know, the greatest kind of bioterrorist.

42:10Ashish Jha: Like nature is constantly running experiments. And the idea that humans can outdo nature on this seems extremely unlikely. And I think Tony was very skeptical of the ability of people to build bioweapons. And Tony's a very smart guy, so I take that with some seriousness. Okay. What is really difficult about this? There are several things, right? I mean, one is you are trying to, in some ways, outcompete nature. You're trying to manipulate—that's why I said earlier, it's very, very hard to build biological agents that are more transmissible because essentially evolution, you know, selects for transmissibility.

42:45Ashish Jha: It is easier to select for things or build things that are more deadly. And, but you have to have a pretty deep understanding of how to do that and how to make sure that those things still continue to replicate. You can build a super deadly virus that kills one person, but so quickly that that person doesn't spread it. You've killed one person, you don't have a bioweapon. So there are still these sort of technical issues that are there, to going back to your cyber experience or analogy, because you needed large numbers of people working on this, it allowed our intelligence agencies to know who they were, to track them, to understand what was happening.

43:26Ashish Jha: The teams have gotten much smaller. And I think from the conversations I've been having with people who work on the sort of biological frontier of this, they feel like the barriers are falling. To pick up on the Tony Fauci theory here, knowing nothing about the history of bioweapons,

43:49Derek Thompson: what are the cases in the last, I'm guessing not more than 100 years, when something that could plausibly be called a bioweapon, a human manipulated virus, bacteria, pathogen was used to kill not one person, but many, many people. Like what are the canonical cases of bioweapons being successfully grown and deployed to kill at mass scale?

44:22Ashish Jha: Yeah, that's a, that's a very good question. So there have been, obviously, most of the stuff has been, it's been hard to nail down, you know, which ones exactly were. I mean, actually think, for instance, in Russia, there was from a lab in, I think, 1979, a release of anthrax that at least many people at that time thought was manipulated, that killed 60, 70 people. The Russian government, of course, Soviet government denied it until the Soviet Union fell. So that comes to mind. I mean, there have been experiments inside the United States that have remained classified where there was work happening on agents, again, not bioweapons, where there were lab accidents and people died.

45:16Ashish Jha: Um, I, I, there aren't very many throughout history of large scale, you know, thousands and thousands of people dying from bioweapons. Um, but I think there are enough in, in history that people look at those instances and situations and worry, um, that the barriers that have kept them small and kept the outbreak small are, are dissolving. Yeah.

45:40Derek Thompson: Yeah, just in a quick search, it seems like, for example, around World War II, Japan targeted Chinese cities by dropping ceramic bombs that were filled with plague-infected fleas, and that this might have been one of the cases where biology was used in order to kill the most people. But that clearly isn't manipulated biology. You're putting fleas already infected with a naturally existing plague inside of a bomb and then dropping it on people. horrendous, but not a manipulated bioweapon. How many biologists that you've spoken to agree with, take seriously the Fauci thesis here, which is that, yes, of course, we should worry about bioweapons.

46:24Derek Thompson: We should worry about people trying to kill people. That's a good thing to worry about. But bioweapons, manipulated biology, is so much harder than most people think that the idea that it's a top order risk to the world can be overrated because it's just so hard evolutionarily to produce a virus that is, as you said, both highly transmissible and highly fatal so that you are killing people, but after a lag so that they infect many people before they pass away. What is the possibility that Fauci's right, that bioweapons really are, actually successful bioweapons are much, much harder to build and deploy than many people think?

47:14Yeah.

47:16Ashish Jha: If you go back and whether you use the Japanese exam for World War II, there's like, you know, there have been a lot of cults. Om Shurikyo used, I think it was an attenuated anthrax, obviously the anthrax attacks in the United States. None of them were really manipulated in a way because it was very hard to manipulate things, right? That's the biological engineering has become a reality only in the last 10 years. So what I say to Tony is you're absolutely right that throughout history that has been extraordinarily hard, but two things have changed. And it's sort of like, you know, again, that just because it hasn't been true in the past doesn't mean it will not be in the future.

47:56Ashish Jha: So again, I think if you ask Tony Fauci, he will, I suspect, continue to say it's extremely hard. I'm not saying it's easy, by the way, and I'll come back to that in a second. Again, there are, I think, kind of traditional virologists who look at this and say, boy, this would be very, very difficult to do. The two counterpoints I would make are, one, this issue that biological engineering is getting easier and easier. Look, 10 years ago, if you had said to me, we're going to take cancer cells out of somebody's tumor, and we're going to build an mRNA vaccine specifically to that cancer. Sounds a little science fiction-y there, Derek.

48:32Ashish Jha: We can now, right? Not perfectly, not every time, but we can now.

48:37Derek Thompson: I believe you're referring to the Balachandran lab. We've had him on the show. Yeah, it's really amazing. These are secondary vaccines using mRNA to prevent the recurrence of pancreatic cancer or other cancers.

48:46Ashish Jha: Exactly. Exactly. It's unbelievable. That kind of technology, or if you think about, look at all the biotech companies that are working on cancer therapy, so much of it is like, can we target the cancer cell, but leave the healthy cell alone? That set of capabilities just did not exist. And so it was true that in that world, it was always going to be mother nature that was going to win. In this world where tools have become ubiquitous, the ability to synthesize DNA has become ubiquitous. The thing that was really missing in my mind was know-how. That you now have tools, you now have materials, you know, synthetic biology, and now finally with AI, you have the third missing piece, which is know-how.

49:33Ashish Jha: You put those together, might it be that we will not see the deployment and use of biological agents against a human population in the next five years, yeah, that could happen. Boy, that would be very lucky.

49:49Derek Thompson: Two more points to make about AI before we get to the question of what do we do about this. I think it's important to say here that the U.S. frontier labs, OpenAI and Anthropic, often have very strict limits to what kind of prompts you can successfully give them to get back any information about biology. Dario Amadei, Anthropic, has said in many occasions, I would much rather deny a science student the ability to learn something about the pancreas than accidentally be the cause of someone learning how to create some pancreatic infection at scale as a bioweapon. So they are, according to some people, overcautious in shutting down certain prompts of biology.

50:32Derek Thompson: But that's not true for all of the open weight models. And I believe you mentioned to me in a pre-interview that, you know, Kimmy, which comes from the AI company Moonshot, is significantly less restrictive. You can ask the questions, you can change the weights, and it can give you answers that are, in some cases, right at the level of sophistication of the up-to-date Claude and Chachapiti models. But in addition to that, in an essay that you wrote in Stat, you cited a study where researchers found that they could successfully order the raw material in order to create some kind of bioweapon. And this, I think, is particularly important that we're combining, what's frightening here is we're combining open weight expertise and in some cases, not enough restrictions around the ability to order the raw materials of viruses and bacteria.

51:19Derek Thompson: Tell me a little bit about that study that you wrote about in STAT and why it's so important.

51:24Ashish Jha: Yeah, and by the way, that was done by Kevin Esvelt, who you mentioned earlier from MIT. So what Kevin did, and by the way, what he did, he actually informed the FBI, the Boston office of the FBI, he's a professor at MIT, that he was going to do this just because he did not want to do anything illegal or at least wanted that. And then he ordered snippets of the 1918 influenza virus. This is what's called a select agent, meaning you can't build it. It's illegal to build it. If we see a version of it, it's it's all you know, it brings law enforcement in. And essentially, he took the virus, split into 38 pieces and ordered it up from companies around the world.

52:03Ashish Jha: He actually went one step further. He ordered it to post office boxes. He ordered it with fake names, things that you would think would trigger concern. Like, it's one thing if it's Professor Kevin Esvelt at the Massachusetts Institute of Technology. And 36 of the 38 orders were filled. And he wrote about it. Obviously, he didn't do anything dangerous with it. And that was like three years ago. Okay. And so now there's a whole push to like, we need more restrictions on the lab. And I think that's fine. But this stuff has gotten global. Like you can order stuff from India. You can order stuff from China.

52:40Ashish Jha: Are we going to restrict and really tightly monitor what? And by the way, those are American scientists. Like what about scientists working in other countries? How do we restrict their ability to get access to this? I just think my view is people are still talking about restrictions, screening. I'm fine with that. To me, the cat is mostly out of the bag here. We just, we're not going to be able to keep this under control.

53:04Derek Thompson: Right. The fear essentially, just to recapitulate here, the fear is not just that someone at MIT does this. It's that someone in Kazakhstan could do this. It's that someone in, let's say, the state of Georgia, in order to create a bioweapon to hurt the Russians that they're in some kind of interstate conflict with, unleash something that then becomes a global pandemic, that you could have sort of interstate relations that have nothing to do with the U.S. that nonetheless pose a threat to American families because of the nature of transmissibility and the nature of viruses and these kind of bacteria.

53:43Derek Thompson: I think that that leads us to the final category here, which is, you know, not just how scared should we be, but what should we do? Let's say you're back in the White House 2029. President gives you, you know, 10 to 20 billion dollars to protect America against this new age of AI enabled biological weapons. What are the first three things you do?

54:06Ashish Jha: Well, actually, I will say when I was in the White House, we started worrying about this. And I talked to President Biden about protecting America. Now, again, my level of concern was lower then, because now we're talking about 2023 into 24, where the AI risks were not quite as substantial as they are today. And basically, we laid out a three-part plan. I mean, one was we needed to have coordination on this topic from the White House. We just needed serious senior people sitting in the White House working on this because agencies are great, but I think they're not going to be enough to really make this a high enough priority.

54:38Ashish Jha: So that's actually part of the whole creation of the Pandemic Office. It was called the Pandemic Office, but it was working on biosecurity every bit as much as natural outbreaks. Second is we built out a plan of what national biodefaction needs to look like. Listen, there are versions of these biological agents that we worry are designed to bypass all of our surveillance capabilities and could spread for weeks or months, could be devastating to the population, and we would not know it until things got very bad. And I can share some of the stuff that we declassified that's actually available in reports.

55:14Ashish Jha: But so we said we would miss those things. We need a national biodefaction strategy. And so we built a plan for that. And then we said we need huge investments in platforms around medical countermeasures, therapies, vaccines. We don't know what's coming. So let's make investments in things that we think are kind of likely, like a platform for respiratory viruses and how we prevent them from becoming, prevent them from not only causing serious illness, but how we prevent them from spreading to other people. And just quickly, what's a platform? Oh, I'm sorry. So platforms, like vaccine platforms are the idea that there's a set of capabilities that you need to develop against for a vaccine that are not very specific to that virus itself.

56:04Ashish Jha: Like, how do you actually deliver the vaccine effectively? So, for instance, we thought a lot about how do you build up mucosal immunity, immunity in your nose and throat, because that's actually going to prevent people from spreading the virus. Remember, one of the problems with COVID is it might prevent you from getting seriously ill, COVID vaccines, but it doesn't really slow down transmission. Can we block transmission with a whole new way of doing vaccines? So a platform would be how do you deliver that kind of vaccine effectively, safely in a large number of people? How do you test it? So there's a whole bunch of scientific work that needs to go in.

56:36Ashish Jha: So we actually made a bunch of investments in that. We made a bunch of investments in the mRNA platform because we thought that was going to be really critical. Unfortunately, three things have happened. The pandemic office sits empty. All those investments were pulled back by Secretary Kennedy, who did not believe that we should be investing in those things. And the national biosurveillance system that we think is actually the linchpin here, because without it, you're just not going to have any idea what happened for weeks, months, until things get very bad has not gotten built at all. And one last thing is in 2025, after President Trump came in, I spent a year working with people inside the Trump administration, Trump White House, DOD, trying to get them energized on this.

57:19Ashish Jha: Actually, it wasn't that hard to get them energized. They very much understand the problem, very much understand what the solutions are. Their ability to effectuate change has been very limited. They have not been able to move the ball. So our country sits pretty vulnerable to this set of risks, despite the fact that we actually know what to do at this moment.

57:38Derek Thompson: Finally, talk to me as a dad. How much should I worry about bioweapons the next decades of mine, my family's life? And what should a family like mine do to prepare? Yeah. Well, I am a dad and I decided a year ago that the chances that a

58:02Ashish Jha: bioweapon is going to end up being used against the human population, as I said, in the next five years is very, very high. And if it ever ends up affecting my family, I need to be able to look at my kids in the eye and say, look, I knew about this and I did everything I could to try to prevent this. So there's a little bit of like an obligation that all of us have to say, if we believe this threat is as real as I am claiming it is, then all of us have an obligation. And I understand this is not quite the question you're asking, but I'll say it like as citizens to like really push our policymakers to look for solutions.

58:36Ashish Jha: It doesn't have to only come from the public sector and their private sector solutions. There are lots of ways of driving this ship forward, and we have got to I'll push on that. From a personal point of view, you know, people ask, are you stocking up on PPEs? I am not. Like, we have some masks at home. I think that's a good idea, but I don't know that there's a ton I can do on a personal level because I don't really have enough insight. Like, we're not building a bunker. We're not, we don't have UV, far UV in our home. Like, there are these things that people talk about as like family protection.

59:10Ashish Jha: And the reason I haven't gone down that road. It's not because I don't love my family. But my view is, you know, I have two kids in college. They're going to be out and about doing stuff. The chances that everybody's going to be at home hunker down seems very unlikely to me. And so this is it's a pretty substantial threat from the outside. We really need a functioning government and we really need a very effective private sector working hand in hand to try to solve this problem. I don't think there's a ton I can do personally. I'm not stockpiling on antibiotics. I'm not like there's just none of that stuff because i don't think um i am smart enough to predict exactly what i need

59:45Derek Thompson: to stockpile in order to keep our family safe just one just one last follow up here like what's the case against the heavy duty masks the uv thing and stockpiling antibiotics like why why not why

1:00:00Ashish Jha: not yeah now that you say it derek i probably should stockpile some masks like you know i haven't. There's no case against it. Like we probably, that probably would be a good thing.

1:00:10Derek Thompson: I'm thinking of the masks that look like something that like someone was wearing in like World War I, except updated for like a century later. Not the thin little mask that someone can wear to like, you know, buy pears at the farmer's market.

1:00:20Ashish Jha: Yeah, yeah. No, and look, Kevin Asphalt, we've talked about, has actually been working on trying to build the next generation of masks and get it stockpiled by private individuals because again, they don't think that the government is doing enough on this. So there are these kinds of strategies. Those masks really don't, there aren't very many of them. And my view is we probably should have some more N95s at home and they're pretty easily available. So maybe we will do that. There's really no good case against it. I have just thought it's a pretty, I'm not convinced it's going to really bail us out if we don't fix the big problems.

1:00:54Ashish Jha: When I think about like the latent virus that we know people have been trying to work on, which is basically a mild cold virus, but it's designed to knock out your P53 gene and essentially cause like cancer, widespread cancer in a population, you know, there's gonna be a cold that will go around. We will never see it. We will not be, I will not be using a mask because we won't know that it's even spreading. Like there are all of these scenarios that I look at and I think, sure, like on the margins, we should probably have some masks at but it really feels like marginal compared to the real solutions that we've got to implement.

1:01:32Derek Thompson: Well, that's terrifying. Sorry. And I really don't like ending on a note of common cold engineered to cause mass cancer. Is there some other note that we could end on?

1:01:49Ashish Jha: Yes, yes.

1:01:50Derek Thompson: Let me actually end on a much better note.

1:01:51Ashish Jha: Because as you might imagine, I'm a blast at dinner parties. Like people don't even ask me like what I think about these things. I could guess. Yeah. But here's the best part of all this. All of this is like, all of this is stuff we can solve. We know how to detect that terrible engineered cold virus. We have that capability today. We know how to build countermeasures against it. The one thing I will tell people that they should be hopeful is there's a lot of activity happening by companies, by nonprofits. in this space, my saying the federal government has largely been absent is not to say no one's doing it.

1:02:28Ashish Jha: There are other countries that are very involved in this. There are countries that are building kind of biodetection capabilities. I've been in close contact with a lot of friends in the UK and EU. There's some really good work happening there. I actually think we can solve this. It's just going to take all of us. And the critical error we could make is to say, ah, the government has to solve it. The government can't or isn't going to or they're incompetent or they're not focused enough. We're all screwed. That is not true. There are lots of good solutions here. And this is what I spent all of my time thinking about and and trying to build for America and the world.

1:03:03Ashish Jha: And I would argue that all of us could be more engaged in being part of the solution. Look, throughout history, we've had governments that have not functioned as effectively as we wish, there are other institutions that can step up, step in and do really great work.

1:03:17Derek Thompson: See how easy that was? End with optimism, end with agency. This is what you need to say, at least during the dessert course at the dinner parties. I try, man. You don't want to leave people with the bad taste in their mouth. People just don't invite me back. Thank you very much.

1:03:30Ashish Jha: Thank you so much, Derek. It was great.

1:03:42Thank you.