Dr. Christopher J.L. Murray — health economist at the University of Washington, physician, and founder of the Global Burden of Disease framework — joins The Future of Medicine to discuss what decades of data across 204 countries reveal about human health, and where it's headed by the end of this century.
Dr. Murray's Global Burden of Disease project involves 19,000 collaborators in 167 countries and has produced more than 5,000 published papers. In this conversation, he explains what that data shows: remarkable progress on childhood infectious disease and HIV, a profound global shift toward chronic conditions like heart disease, diabetes, and chronic kidney disease, and a mental health crisis — rising steadily for over a decade — whose causes experts still actively debate.
He also discusses why the widely circulated claim that American obesity rates had flattened is likely a statistical artifact, what the data projects for the coming decades (including diabetes overtaking heart disease as the world's leading cause of death), and the compounding health crisis facing the Sahel as population growth, poverty, and climate change converge.
The conversation closes on the Lancet Commission on existential and catastrophic risks — a project Dr. Murray helped initiate that takes a data-driven look at long-range threats to humanity, from AI and synthetic biology to nuclear risk, out to 2100.
It's a rare conversation that takes a long, honest look at the future — and doesn't flinch.
The Transcript
Dr. Euan Ashley: Chris, welcome to Stanford.
Dr. Christopher Murray: Great to be here.
Dr. Ashley: It's uh we just heard a wonderful grand rounds. You spoke to a packed room full of people, standing room only. Uh and so now we have an opportunity to chat a little bit, uh about uh the global burden of disease, about your career. Uh but I wanted to start at the beginning like and I love it when we have guests here and we have a chance to chat a little bit about how you ended up where you are and and so my question is this. Did you ever as a kid imagine you were going to be running this, you know, this enormously impactful study from the position you're in? And how did you end up with so many uh degrees and qualifications?
Dr. Murray: Boy, uh simple questions, right? Uh, my parents did medical work in Africa. So uh my dad was a cardiologist. Uh were from New Zealand, moved to the US for work, he did, at Mayo clinic and then University of Minnesota. And then at some point in the 70s, yeah, you know, a lot of turmoil, uh my brother's a little bit older said, you know, what are you doing for society? You're just, you know, this. And so to respond to that, he took a sabbatical and uh we bought a couple of Land Rovers at Sully Hall in England.
Dr. Ashley: Wow.
Dr. Murray: Arranged through our church to find a hospital built in eastern Nishair that had never started, and we drove out and my dad used the kids and the family as the labor in the hospital. So at 10, I was my dad's pharmacist and surgical assistant.
Dr. Ashley: Uh and so you you drove from from England.
Dr. Murray: We drove from England. We went through Spain, Morocco, across Algeria, across the Sahara and got to this place.
Dr. Ashley: Wow.
Dr. Murray: And then we ran this hospital.
Dr. Ashley: And you were 10 years old.
Dr. Murray: I was 10. And then my parents were hooked and they kept doing that for the rest of their careers.
Dr. Ashley: But even prior to 10, you were already, you spent a year, some years in New Zealand and then sometime in Minnesota.
Dr. Murray: Yeah. Some time in New Zealand. My dad had done one prior sabbatical in in Oxford. So I went I did a little bit of schooling in Oxford as well.
Dr. Ashley: We'll come back to that.
Dr. Murray: So, you know, we were sort of a already a little bit of a globalist family.
Dr. Ashley: Yeah. So, you were working in a hospital in Africa at 10 years old helping your dad.
Dr. Murray: Yeah. And and then they did work in uh the Camaro Islands in in the Ogaden of Ethiopia. A lot of years in the Rift Valley in Kenya and then their final stint was in Malawi by which time I was off, you know, doing school and uh university. So, that set me and, on a path.
Dr. Ashley: Did you ever consider doing anything else or was it always going to be medicine from an early age?
Dr. Murray: You know, I went through a crisis of what should I do with myself, you know, like when I was a senior at Harvard and the choice was sort of something in global health because I wanted to have more population impact than I'd seen my parents having. Yeah. Uh but the other one was I really love the outdoors and animals and so I thought about being a park ranger in Africa, so that was the big choice. Okay.
Dr. Ashley: Well, let's you know being a doctor to humans or on population scale or doctor to animals on a slightly smaller scale. Those two things align but in the end you chose humans.
Dr. Murray: I chose humans. I got a fellowship to go to Oxford and a Rhodes scholarship. I got a Rhodes scholarship and uh that was great because then I found the right person to work with who was not at Oxford, it was at the London School of Economics, Brian Abel Smith. Yeah. So I went down the I want to have more impact so I need to learn about money and how health systems work and I so I did health economics and then I went back to medicine at Harvard after that.
Dr. Ashley: So even before you finished your medical degree, you'd already taken a kind of side turn towards economics.
Dr. Murray: I had. Yes. Yes. And I, you know, partly this sort of quest, which I'm still on, uh, how do you like make largescale change and got stuck along the way well, we don't know a lot about what people's problems are, right? And then, you know, we really should try to fill in the gaps, know more, figure out what what works and and how you can have impact. And, you know, that's the journey.
Dr. Ashley: Well, I was just talking stimulated by your talk actually just talking to one of my colleagues about this idea that we're interested and we're interested certainly here at Stanford in having a big impact on the world and on the world of medicine and the world of health. But the starting point for that is understanding what are the big problems. And there really is no better resource. And obviously before you built it, there wasn't really a resource to understand what are the biggest problems in health uh today across the world as we'll come to talk about. You're also projecting forward to the future to help us think about what the big problems might be for humanity going forward. But uh in your case it sounds like that you saw the need and you were you were driven by the the the desire to make change and realized that you didn't know where to kind of fire that cannon. So that's what led you to start to quantify diseases. Is that what it was?
Dr. Murray: That's sort of why that question about you know what are the big problems was so uh you know preeminent in my mind and then I just got this great opportunity which was when I came back from Oxford and was in med school I was I was also a fellow at the Harvard School of Public Health in Lincoln Chen's department at the time, and somehow I I don't even remember the sequence of events but I started doing some work with the World Bank. And when they came round to deciding to do their annual policy assessment on health for the first time, uh fortunately for me the person running that team, Dean Jameson, uh you know turned asked me if I could help with and then with Alan Lopez at WHO answer the question and turned out that that was a non-trivial undertaking.
Dr. Ashley: Right. Right.
Dr. Murray: Uh but you know we really did back of the envelope pretty simplistic stuff with what we had at hand in 1991-92 but that just got you know got some momentum and just sort of kept going and then I went off to WHO with Grow Harlem Bruntland to do it there. And then she stopped uh didn't do a second term as director general and that's where I sort of moved back to academia to do it.
Dr. Ashley: Yeah. Interesting. Jump us forward now then to the latest update and where we are and we can sort of fill in the the parts along the way. But the impact of your study, it's hard to even call it a study. It's much more than that, but give us some statistics to give the people listening here just a sense of the impact and the size, shape, and impact of the global burden of disease.
Dr. Murray: You know, it because it's just kept going and and growing. Uh we have a ton of people involved. So we have 19,000 collaborators –
Dr. Ashley: 19,000.
Dr. Murray: 19,000 in 167 countries, uh intentionally like we we target countries that are not active to find people because we want it to be truly global right, uh but because of that then lots of academic output you know 5,000 plus papers from the GBD, uh but probably more importantly just tons of use by funders and governments as the sort of first stop to sort of understand how big is disease X or problem Y.
Dr. Ashley: Yeah. And well, you talked I mean you you talk about 5,000 plus economic papers. First of all, those were in mostly in pretty big high impact journals.
Dr. Murray: A lot of them are in good journals.
Dr. Ashley: Yeah. Um but also you showed a slide at the talk just now where I think you'd mentioned you you'd used some agentic AI systems. We'll also come back to those, uh to try and sort of measure the impact on policy. Tell us a bit more about --
Dr. Murray: Yeah. I mean, because we've always wondered like who's and we get these anecdotes, you know, the the president of Batswana used the alcohol numbers to put through legislation on alcohol type story and we, you know, replicated, but, you know, we we're data people, so we wanted to have a way to see if we could track this a little better too. Nobody wanted to pay for it. But now with agents, you know, we just sent them on their merry way and said, "Go find uh all the documents on government websites that make use of the GBD." And then once they collected them all, we said “read all these documents.” There were like 4,000 of them. Uh and make a typology of uses both by topic but also by like what what's the purpose they're using. Is it for resource allocation? Is it for priority setting? is it you know just general strategy development and it did a nice job of that. So what we realize is there's a lot of uses. Some of them we never even thought about. Like for example now a hot topic is using the forecasts to predict how many specialists in different clinical specialties we should train. Because you want to train not for today. You want to train for 20 years from now because they have a 40-year career, let's say. And so you know, we didn't come up with this use case, but people find clever ways to to use data and forecasts.
Dr. Ashley: Yeah. And I think that the testament to you sharing so broadly, uh not just through publications but also allowing the data to be available for for collaborators to to work on it.
Dr. Murray: Yeah. And for anybody to go to the website you know GBD compare and and look and explore right.
Dr. Ashley: Yeah. So the thing you must get asked all the time I'm sure at parties and everyone else is you know is you know what are the headline findings? I mean what are the things that surprised you? What are the things that have changed? And you had some nice slides summarizing them. So I know that you're used to being, you know, having to find a succinct way of summarizing. I mean, I I would, you know, half the room here were taking photos because the ones who don't immediately know which paper to go to to see those, you know, that that's not information that many of us carry around in our heads. And yet probably it should be.
Dr. Murray: So yeah, you know, I we do because there's so many numbers and so much detail, it's sometimes easy to even for us who are intimately involved to just get lost. Yeah. So we do it takes a bit of doing to say, okay, let's boil this down to some high level messages, right? And and you know, so we try, uh and I'm sure somebody coming from this is my favorite problem will say, well, you left my problem out. With that caveat in mind, you know, at the macro macro level, the world because of population aging, because of progress on infectious diseases in kids, particularly because of ART for HIV, uh we've seen this big shift from the infectious diseases, the maternal causes, the nutritional deficiencies, towards the non-communicable diseases. Really profound. And even in the low-income world a little bit, but very dramatic changes in a in a place like Indonesia or Brazil or India even where it's the pendulum is really swung over to to heart disease, chronic respiratory disease, and then NCDS more broadly mental health um you know the muscular scaladal disorders those also making h you know being a big factor.
Dr. Ashley: And I think for some of the people listening even that terminology would might not familiar to them. But we we define non-communicable disease almost like the exclusion, it's the inverse of communicable diseases which were infectious diseases because those were the biggest killers when when you started this, right?
Dr. Murray: Yeah. And in fact, this sort of it's a weird label for sure. And the non-communicable diseases just become a thing globally, right? You know, there's UN highlevel summits on NCDs, right? But yes, it is it's the heart disease, cancer, diabetes, chronic kidney disease --
Dr. Ashley: Chronic degenerative diseases that people won't be very familiar with.
Dr. Murray: Right. Right. Uh so that's one really, really macro view. And then if you drill down, you know, a step, we've had this incredible progress in reducing death under age five. Yeah. Less progress on deaths in adolescence, but the rates are much lower. Okay. But still tons of progress but still huge disparity. So it's like still 10 to one on the risk of dying under age 20 from the low income world to the high income world despite all this progress we've had.
Dr. Ashley: But the curves are parallel each other. So that there's improvements in both places.
Dr. Murray: There's improvements everywhere but the improvements are bigger in the low- income world. And part of that has been the success well up to now of the money that's been spent by the high-income world on programs in the developing world right. Those have actually been really successful: diarrheas, pneumonia, malaria, super super impactful uh HIV yeah another huge impact. So you know that that's one high level message. So success but huge disparity still disparity still. uh Secondly, you know, that we talked about the big shift to NCDs, but it's really strong in the sort of Asian countries in Latin America. Yeah. Uh third is the rise in mental health as a as a we're seeing in the data there was a there's still a lot of debate like we're now seeing you know well-conducted surveys using the same methodology that show a big increase in anxiety and depression. Yeah. Why is a debate.
Dr. Ashley: Yeah. Not just around Covid though.
Dr. Murray: No, not just around Covid.
Dr. Ashley: Because that seemed clear at the time. I mean that was widely reported.
Dr. Murray: No, it started well before, probably a decade before and it's trending up. Big spike during Covid but still going up and uh now we get into really heated debates is the why. Is it Jonathan Haidt's social media, right? Is it, you know, devices? Is it despair about you know employment and you know the future? Is it climate change despair? Long list. And when you talk to the mental health epidemiologists they do not agree with each other. Right. So there isn't a consensus in that field right on the why. There's now a consensus that it's getting worse. It is getting worse. So that's that's a really interesting phenomenon going on. Uh and then of course the obesity epidemic is getting worse everywhere. Yeah. Not just in the US. In fact, the worst in the world is sort of Mexico, the Pacific Islands, right, the Middle East, uh, are the places where it's it's the worst. Uh, but it's just going up everywhere.
Dr. Ashley: And but you alluded to this just briefly in in your your talk. I mean, there was a fairly widely um distributed graph of America's obesity rate that had kind of leveled off a little bit. It gained quite a lot of traction last year. I can't actually remember it its origin. You kind of alluded to it and said that there maybe there had been some walking back of of that.
Dr. Murray: Well, remember each round that what that was it I think it got picked up in the Financial Times actually. It got spread everywhere because that uh but what that is that each round of NHANES in this country is a pretty small sample and you it's sampling a very small number of communities because they have to drive the big bus vans around to do the measurements. Meaning that the uncertainty intervals on the measurements of obesity are like plus 2%, minus 2% something in that range right and so there isn't actually a statistically meaningful flattening yet and when you look at the self-reported data yeah which obviously not as good but still tells you about trends it's not flattening yet so that's why I think we're sort of skeptical. Now GLP1's should
Dr. Ashley: I mean there are 50 billion dollars worth of uh of spending on those drugs. I think almost everyone can point to not just one but probably multiple people they know who have changed their appearance dramatically.
Dr. Murray: Fabulously impactful, but seems like at some point that will right and especially once the full you know the fully oral ones are being marketed
Dr. Ashley: Which is just coming right now basically.
Dr. Murray: And then so I'm I'm pretty hopeful that we'll see some breaks on that. Yeah.
Dr. Ashley: Uh and this might be the beginning of a signal, but that particular signal --
Dr. Murray: That particularly one was oversold. Yeah. Got it. Got it. Yeah.
Dr. Ashley: So these are these big categories and you've mentioned how they've changed over time. You mentioned the the disparities. What what what's give us a sense of the size of them at the moment. You have this great graphic. You you talked about the rectangular pie chart. Yeah.
Dr. Murray: The tree map.
Dr. Ashley: You know, what are what are the big uh health signals there or disease signals there that uh that we can explain to the the people listening.
Dr. Murray: Well, you know, the ones that have always been there and that's why there's so many cardiologists out there have been eschemic heart disease and stroke, right? Yes. Uh and and you know, to a lesser extent some of the other uh you know, cardiomyopathies, etc. Uh but now we see you know really diabetes and chronic kidney disease on the on the rise at the global level. those are becoming very prominent in terms of within that bucket of the non-communicable diseases, right? Uh lots of mental health and depression, anxiety. Um we're seeing, you know, the the bucket of back pain, you know, neck pain, all the causes of those uh that are unfortunately we don't have a lot to offer for them and they go up with age. And so as the world ages, there's just more and more and more of those. Yeah. Um, we have, you know, dementia going up in terms of numbers, not the rates. Yeah. But the numbers of people. And then on the injury side, there's an interesting one that we didn't talk about too much in that presentation, but falls are on the rise. And there's all sorts of theories as to why, you know, more elderly people living on their own, you know, all sorts of things that might be accounting for this pretty pretty substantial effect of falls. Some age groups, you know, linked to the mental health suicides are up in some of the adolescent age groups. Um, and then of course we have the ongoing malaria. TB, pneumonia, diarrhea, neonatal causes. Yeah.
Dr. Ashley: I noticed and you you showed this contest between I think the 19 early 1990s data and the 2023 data. Uh, and of course there's changes you've discussed infectious diseases, but it's hard to not notice measles in the middle, which at this stage is smaller, but we're all aware of the data that it's
Dr. Murray: well, measles was a top 10 cause in the world in 1990. It's now like 70th or 66. So, you know, we've had because of vaccination. Yeah. Huge impacts on measles. Now, of course, we're in a period where vaccination rates have dropped in a number of of geographies. Fortunately, not so much in low resource settings yet. Uh but you have to worry, right? Uh because of the funding for for programs like the global alliance for vaccines or copy.
Dr. Ashley: I mean you talked about the numbers of people who've used your data. I imagine you know as the encyclopedia essentially of health now and with some projections in in the future and you must get a lot of interesting incoming questions. I'm imagining from legislators but also I assume from pharmaceutical companies and others who who are interested in where they should apply their R&D. Tell us a little bit about about that side or tell us some interesting stories from that part of the of your life.
Dr. Murray: Yeah, I mean pharmaceutical industry is interesting right because they really are uh avid users of data. They are they are very much a data driven industry right which is good right because that's what we want. So you know they and it's both a strategic use like where should our where should we put our capital in terms of you know new products right and you know new markets for for existing you know should we uh you know Takada has a dengue vaccine you know who should be the next place they target for licensure you know that sort of type question so we have a lot of engagement uh with you know from or questions that come from pharmaceutical industry they they have their own in-house epi expertise as well. So you know they look at our stuff and then they'll you know try to you know verify that their own in-house analyses agree because they make decisions that are quite costly.
Dr. Ashley: Right. Billion dollar investments development.
Dr. Murray: So that that's one group. You know, we see actually quite a bit of uh use by um some people like in the healthcare delivery side of things, not just pharmaceuticals. uh they're they're interested in are there uh conditions for the groups that they're looking at that they haven't targeted yet or right and then there was a period still some out there where people are interested in plugging in the the epi into automatic diagnostic you know bayesian diagnostic algorithms you know like uh you come in with a certain symptom cluster and if you're in Rwanda it's probably a different disease than if you're in Cleveland and so we can just they can take what we have and feed that into --
Dr. Ashley: Insurance companies presumably as well.
Dr. Murray: Insurance companies some um you know they're very they're often more interested in death than in the cause of death or the disease. There's some some exceptions to that. Yeah. uh reinsurance people somewhat, you know. We've started to do a lot of climate modeling and work to for by necessity. Right. And so then some of that starts to be interesting like you know where where are storms and floods and things like that.
Dr. Ashley: Well, you'd also talked about projecting into the future. So I'd love to talk for just a minute first of all about the the disease boxes that we were talking about in a moment and what they look like in the future and then we'll get to some of the really big questions. So let's start with there. If you project that box kind of your your rectangular pie chart tree uh chart into the future what what what are the significant changes for the next 10, 20, 50 you go right out to the end.
Dr. Murray: We go right out with with trepidation 2100, so you know the the continued trend towards more NCDs but a because of obesity a little bit in interaction with temperature uh really big increases of of diabetes and chronic kidney disease so that diabetes becomes number one. So it super it supersedes eschemic heart disease and stroke is at the top of the list. Uh we see you know ongoing more mental disorders in the future. uh stubborn uh communicable you know the infectious disease problems in places like the Sahel because there's so many more people in the Sahel as you go through the century because of high population growth. Right.
Dr. Ashley: You mentioned that although the frequency of some things may go down because it's out um played by the population growth so the absolute numbers go up.
Dr. Murray: Yeah. I mean take if you want a concrete example uh take Niger right which has had it barely has changed its income per capita in 75 years, hardly different than 1950. Yeah. gone from a child death rate of like 350 out of a thousand babies dying before the fifth birthday down to about a hundred. Yeah. Maybe slightly more. Yeah. So huge progress in kids, little bit of progress in adults. Government's not particularly functionable. So it's this is really a testament to the programs that are being funded externally and delivered uh you know with with obviously local people delivering the the interventions. Uh but they have not really changed fertility very much. So the still most women have six kids. You put that into a place that's just a thin strip of Niger's habitable because the rest is the Sahara and and there's desertification. Uh and you run that into the future, you get the population of Niger just going up by many multiples, right? And so the these are the population growth is in the places where people will likely stay quite poor likely have the biggest effects of climate change. Uh so you get this sort of concentration of lots of health problems uh in the Sahel belt all the way across. And that's why despite general progress, we're still going to have malaria and pneumonia and diarrhea and neonatal disorders even as we go out into the later in the century.
Dr. Ashley: And you project late? Yeah. Later in the century, we're still thinking of those as significant.
Dr. Murray: Which, which surprised me, right? I thought, you know, we've made so much progress in 75 years that the people who've been left behind somewhat even though they've had quite a lot of progress, would just sort of grow out of it. Which is why I keep coming back to what can we as the world do to help accelerate economic growth in these places because without that you it's hard to see you know this this story of this this gap with the Sahel and the rest of the world not not being a really big issue.
Dr. Ashley: Yeah. Yeah. Amazing. There are very few people who I think are really thinking within concrete terms with data this far out. I mean we're usually talking five or 10 years.
Dr. Murray: Maybe we shouldn't be. Yeah.
Dr. Ashley: Well, that that's maybe what we'll come to now because uh this one of the interesting places that this study and your career, your work has has drawn you is into very large questions and particularly uh the focus of the Lancet Commission um includes everything up to and including human extinction. So, so tell us a little about that, how it came about and you know what you have been concluding recently.
Dr. Murray: So I guess three and a bit years ago I started talking to Richard Horton the editor of the Lancet and to Natalia Kanem the head of the UN fund for population activities about the fact that uh we had we were still in the pandemic but coming out there was quite a lot about climate but were we sure that and and actually Ukraine had just been invaded right uh were we sure that was it? Is were those the three things we should worry about and we can not worry about lots of other things and you know we had the idea that uh maybe we should have a a data driven evidence-driven approach to thinking about what are the big problems. Yeah. And then we had to have a you know we had an interesting group come together as commissioners and then we had some really interesting almost philosophical debates like is it reasonable to look out to 2100 and then the climate people all come and say of course because the carbon we put out in the atmosphere is going to have effects for the next two 300 years. So, of course, you have to look out far because that's where the harm is. And then other things, you know, once you take a longer term perspective, you you also start to change your view on on some other threats, right? So, we said, "Yes, okay, we'll go up to 2100. It's longer than most people are comfortable, but the climate folks have already made that leap for us." Uh, for good reason. And then we started digging into what could be the set of threats and trying to find a way to look in a very even-handed way uh and both the ones that come straight out of the burden of disease type assessment but all the other ones that people are worried about you know uh obviously AI uh you know nuclear winter and and on down the list and including even in the discussion we just had somebody came up to me afterwards said you should have kept put in mirror life.
Dr. Ashley: Right. Well, let's let's unpack that for a moment because certainly on this podcast, we've talked quite a lot about AI. Uh, over the last few months, things have changed dramatically in that sphere. In fact, the first time we met was probably a year ago, I think. And I remember you put AI on the list. And at the time, I was still a bit like, well, that makes sense. But I was still a little skeptical. I've changed my mind completely in the last few months. I've changed most people's minds. I mean, I think um if if you're not paying attention at this moment or if you're not a little scared, you're probably not paying attention and so it's completely reasonable. Um but yeah, talk us through the process that led you to and and maybe explain what mirror life is for the uh for the listeners.
Dr. Murray: Well, I you know, mirror life is this idea that all of evolution has been producing left to right left. Sorry, it's been organic chemistry is a long time ago. Left-handed uh you know molecules and if you make right-handed ones the immune system and others would have no defense.
Dr. Ashley: and now we can with synthetic biology make whatever molecules we want.
Dr. Murray: So so there there is the possibility of really terrible things being made. Uh but how do we come up with that? Well part of it was people putting on the table from the data. Yeah. Putting things like AI because it was already out there that people were worried. You know, you have philosophers at Oxford like Toby Ord writing about the potential catastrophic risk. And there's this group strong here, I think, in the Bay Area, of effective altruists, you know, the EA community who think about thousands, if not millions of years into the future and say, you know, what we should really be worried about is, you know, getting off the planet and this sort of thing. So, we were trying to listen to that, but not get caught up too much in the overly Star Wars type stuff. But you know once you start down that road then you put on the table things like AI and once you look at it particularly AI being used to design highly transmissible lethal pathogens. Because a bioweapon doesn't have to survive natural evol because the the evolutionary argument is oh if it's too lethal it won't continue to spread. The terrorist or the person trying to do harm doesn't care. Yeah. Just wipe everybody out in the first go. Right. Uh so that becomes a very and you know I think all the people that I talk to in close to the the foundation labs are super worried about the bioweapons part uh for for legitimate reason. Then there's other, you know, there's loss of control. There's systemic, you know, manipulation of by AI of systems that can have really quite large Yeah. Just think hospitals, right?
Dr. Ashley: And there's a big debate, I mean, it's literally in the news this week around the defense department.
Dr. Murray: Deputies, killbots. Yeah.
Dr. Ashley: Changing their contracts and whether a line in the sand was was drawn by Anthropic in particular and whether that line was different from from other of the frontier model companies. uh but there is no doubt that that each of those companies and I think in particular Anthropic actually has a section that is concerned and focused on the autonomous element of the AI and the harm it can do and they have regular tests that they run at Anthropic specifically where they test those and they test each new model in that way and those are the people who are worried.
Dr. Murray: Yes, yes it's seems very legitimate to me to be worried and you know from our point of view it's is there a risk answer is obviously. Yes. Does it matter if it's, you know, 1% or 5% or 10% extinction risk? Probably not. It just that there is a real catastrophic risk there. And so, we should take it quite seriously.
Dr. Ashley: And I I mean, in one of my AI talks, I I put up the slide of the P doom, the probability of doom that is is held by a number of the prominent AI researchers and and CEOs and others. But the I mean first of all in almost no case is it zero which is the point that you're making. Uh for some it's low uh you know in the low percent range but for many it's 10 to 20%.
Dr. Murray: It's a big risk.
Dr. Ashley: That that's a big number from people who truly understand how this technology works. Yeah.
Dr. Murray: Yeah. You know and uh there's an interesting divide right when people like between the people who work in those companies or experts always higher then this class of people called super forecasters which these are people that have generally nobody knows why are good at predicting the future and they --
Dr. Ashley: Are you one of those? Aren't you a super forecaster?
Dr. Murray: no I just set up the systems that'll allow the models to predict uh but yeah I I think it would be a grave error to dismiss it. And when we started three years ago, because we already had it on our list at the from the beginning of discussion, there was a lot of giggling. Yeah. You know that there is the uh The Terminator really put us back, right?
Dr. Ashley: I know. But the more you think, the more you worry that there are fiction writers from the past, you know, perhaps were were quite prescient in what might be possible. Um, and I agree when I first made the slide of of Jeff Hinton suggesting that we really should be careful about human extinction, it was partly as a joke. I mean, you know, and and then I contrasted it with multiple other of the the Nobel Prize winners for for AI who were countered to that. But the reality is today that's that's not really a joke anymore. Um, and there's examples of real harm and real harm done at scale. And to your point, I think you don't need to understand much of biology or AI to understand that the possibility of biological warfare is real. Um, so the question is this is really beyond your realm, but what do we do about all of that? I mean this is a question.
Dr. Murray: Well you know on the bioweapons front um there are you know proposed strategies you know more careful regulation of who can buy the the right ingredients you know the nucleotides etc uh and you know monitoring of purchases of anything that could be a precursor. Um, so there's a there's a package of sort of monitoring regulatory stuff that people talk about. I don't know what we it's one of those things you would only know uh that they probably won't tell you if if the government detects this, right? Uh the one that's a little systemic risk, you know, you can try to think of ways to make hospitals and other critical systems more resilient. The one that's really hard is the alignment one because the big labs are trying. Uh but of course the argument that they all make is that the competition with China is such that nobody's really going to slow down.
Dr. Ashley: It really is an arms race at this point.
Dr. Murray: Yeah. And so that's really worrisome. Yeah. Right. Yeah.
Dr. Ashley: No, an interesting we can't we obviously can't end on that note. No, because we have to think about there's a lot of positives in of course from from your study over the years. I mean, I think the fact it exists is a huge positive. I mean, it's just a congratulations. I mean, it's it's an amazing achievement. But but I think uh while you know, a lot of the nuance is only available when you dive a bit deeper as we have in the in the last few minutes. Um the overall trends some of them are incredibly popular when you you talk sorry incredibly positive when you know you talk about the reduction mortality in age out of five. There is progress, right?
Dr. Murray: tons of progress. And you saw the graph of life expectancy that despite climate, despite these risks that are out there, our best prediction for the future or most likely, not if things continue as we're the trajectory we're on is things will continue to get better, right? So, we are fundamentally sort of saying, yeah, we think the world will get better. We can do a better job. We can, we can try to address some of these threats in in a more effective way. But on balance, the future's probably going to be better than the past, right? Um now, you know, the world's full of half half full half, you know, empty type people. Uh but, you know, we're basically saying there there isn't, you know, the other one that's really positive is that climate's a serious risk, but it's not a catastrophic risk of the century. Right. And so, you know, some young people, you know, like at school, they're they're really anxious about the world getting too hot for humans to survive. there won't be enough food. None of that seems possible this century, right? Uh you know the people lots of bad things but it is less extreme than some people who are very anxious about climate may think. So there's some positive messages there. You know we shouldn't go to sleep on that watch.
Dr. Ashley: But no complacency there.
Dr. Murray: No complacency but it's not shouldn't be a source of you know existential angst for kids and others.
Dr. Ashley: Yeah. Well, Chris, thanks for the amazing work you've done over decades in building this this unique resource really for the for the planet. Uh, and thanks so much. This was your first time to Stanford, right?
Dr. Murray: First time to the med school at Stanford.
Dr. Ashley: Okay. Well, we'll have to make sure we have you back sooner rather than later. Thanks so much for joining us.
Dr. Murray: Thanks, that was great.