← LSE: Public lectures and events
LSE: Public lectures and events
Technology for the public interest: preventing capture and promoting welfare
01:27:16
Loading summary
Transcript70 lines
- [00:02]
A
Welcome to the LSE Events podcast by.
- [00:04]
B
The London School of Economics and Political Science.
- [00:07]
A
Get ready to hear from some of.
- [00:08]
B
The most influential international figures in the social sciences.
- [00:17]
C
Hello everybody in person made it through the rain online. Welcome out there. My name is Ken Chadlin. I'm a professor, Department of International Development here at lse. It's a real thrill to have Padmarashi Gelsempath with us tonight at the lse. Padmara Shi is a visiting professor in practice in the Department of International Development. She's also a senior fellow at the Harvard Kennedy School and she wears many hats, I think most relevant for tonight's lecture. Padmarshi is the CEO of the African Pharmaceutical Technology foundation which is based in Kigali. Patraschi has a very long and distinguished career in academia and in international organizations. We figured out that we've known each other for about 20 years now. We first met from a study that she did for the World Health Organization in which she was looking at the changing strategies of Indian pharmaceutical firms in the context of World Trade Organization agreements and the introduction of intellectual property rights in the sector in India. And we've had a long standing sort of series of relationships dialogue over the years. When we first met you were at the United Nations University in the Netherlands. Then she was at UNCTAD in Geneva. Prime Russia is always full of creative insights based on grounded, careful empirical work and really looking forward to the presentation. Tonight she's going to be talking about two technology based sectors that receive a lot of attention in the news and in the but are often not often compared explicitly and this is the pharmaceutical industry and artificial intelligence. After Padmachery finishes, we'll have a discussant comments from Laura Mann who is a colleague of mine in the Department of International Development. Laura works on the political economy of information, knowledge and technology in the global economy. And she's completing a book project on the digitization of global agriculture which is based on research that she's done in the west of the US and the east of Africa. When Padmarashi and Laura have finished, we'll open the floor to questions. Normal people here and those of you who are online, if I could just ask you to please put your phones on silent. That includes yours because it's gone off like a hundred times in the last couple hours. Seriously, to not disrupt things too badly. The evening's event is going to be recorded and will be available as a podcast, assuming that there are no horrible technical glitches which I'm counting on. There won't be. So I'm going to turn it over to Padma Rasheed, your show and then Laura will follow up. Thank you very much. Please join me in welcoming Padma Rashid Alicee.
- [03:29]
A
Good evening everyone. I first want to thank Ken for this very kind introduction and I hope I'll be able to do justice to that. I want to also thank Dr. Laura Mann for agreeing to be a discussant and thanks to all colleagues in LSE International development who made this event happen. I'm going to speak on technology for the public interest and as Ken mentioned, I'm going to be comparing two sectors which are not usually compared, AI and pharmaceuticals. So. All right. Okay. So my presentation will first begin with brief explanation of winner takes all markets, which is sort of a kind of term. Winner takes all is used actually in economic theory for certain kinds of markets. So I'll explain what they are and why it matters. I'll next move on to looking at pharmaceuticals and artificial intelligence, both the similarities and the differences to show how these industries create and maintain advantages. I'll then move on to talk about how pharmaceutical companies mimic winner take advantage. Most models, how do they create those dominant positions and how do they maintain it? And then I'll move on to implications. So what that means really? Oh, okay. So winner takes all explained. So a winner takes all market is where you have markets where best performers disproportionately capture large rewards and market share and leave the competitors with minimal returns despite similar efforts and investments. Winatic all markets have certain specific characteristics. They have network effects. So your value increases exponentially with more users, creating competitive advantages that compound over time and lock out other competitors. Think about platforms, for instance, social media platforms. Easy example. There's economies of scale, so there's large fixed costs that create barriers for entry while marginal costs decrease. Right. And this enables actually dominant players to offer better prices and services. And there's always a first mover advantage where speed matters. So in 2012, for instance, Mark Zuckerberg sent out an email to his team saying, oh, Instagram is attracting users with a very simple feature. It applies for mobile phones. So we immediately need to launch Facebook's own mobile phone app. So who gets to the market first and how they get to the market really matters. So it's either the market. So in winner takes on markets, it's either the market structure itself that alters because of technological advantages, because they amplify these advantages and turn them into dominant market positions. Positions for. For. For some companies, or there are other factors that make it happen. Now. So how does the winner takes all Dynamic look in AI. So what it does is that you have research intelligence which sets the foundation. Then you see, okay, here what I'm doing is I need. Yeah, so as opposed to the old platforms, older platforms, I'm looking at how lock in forces work in the context of AI using hardware ecosystems, software ecosystems and then a lot of investment and incentives which directs all the R and D investment towards scalable AI approaches today. Right? So you have research and talent which sets the foundation. Then you have data modes where you have proprietary data advantages, then you have infrastructure investments which are pretty huge into platforms and ecosystems. And then you have ecosystem adoption. Once you have the ecosystem adoption, you have more users using this ecosystem. And that automatically creates a self reinforcing virtuous cycle which feeds off on the advantages for the company that creates that first. So a good example is Nvidia. Right? Okay, so what happens here is you see actually a hardware foundation, which is the basic layer, you have the moat, which is the software foundation. You have then the network effects and then you have the data flywheel. So you have a virtual cycle where more AI developers drive chip demand, you generate R and D revenue for better performance, and then you actually continue the advantage of the company that created that. Now this is what actually very simplistically put, but actually my presentation today is based on a paper, it's available online, hopefully next week, but here I'm summarizing the AI analysis from the paper. So very simplistically put, you have, this is what defines the kind of AI oligopoly that we see today. So you have cloud giants, you have foundation model leaders, and you have hardware and infrastructure providers. But you see some firms or just one firm which then emerges to become the undisputed leader. And you see a sufficient lag, both innovation and user wise and also in terms of market advantages for the remaining ones. Right? So now in here I'm comparing this, I'm using that as a background in this presentation today to look at what's happening in pharmaceuticals. Now when you compare the two industries, AI and pharmaceuticals, you see on the left side there are some specific aspects that define actually when I take small markets, right? So in the case of AI, you have a core asset, which is your ecosystem, you have network effects which are very strong, okay. And you have barriers to entry, which is data computation, scale, also R and D investments. And then you have, in the pharmaceutical sector, none of these are met. So if you take for instance the core asset, you don't have an ecosystem, you'd normally have a Patent on a molecule or a therapy. If you look at network effects, they're weak to moderate. Okay, so there's no real advantage for, you know, a drug doesn't become more effective the more people that use it. A drug's efficacy remains the same. Right, okay. And then the barriers to entry are not data and computation skill, but they are regulatory approval in clinical trials. Then there's marginal cost for application in a winner takes all market normally is near zero. In the case of pharmaceutical markets, it's, you have high scale economies, so the more you produce, the cheaper it gets. So diminishing returns. Sorry, diminishing, diminishing costs of manufacturing, then the IP strategy is the same. The regulatory environment is also not the same because in the case of AI, it's nascent and evolving and pharmaceuticals, it's mature and stringent. So what exactly explains what's happening? Right, okay, so that's what the paper looks at. Why do these two sectors, which are so different, end up behaving similarly? So the reason why they behave similarly is that in the case of pharmaceuticals, the network effect doesn't exist. But firms create an ecosystem lock in effect, which is similar to that of AI to create a winner takes all effects. The barriers to entry are not a software stack or hardware performance or scale of R and D. But there are the regulatory hurdles, the manufacturing complexity and patent tickets. The primary defense mechanism in the case of AI is technological modes. So it's constant innovation to make your own hardware obsolete. But in the case of pharmaceutical, it's device modes, patent modes, therapy modes, data modes. So it becomes extremely complicated. And so now how does dominance work in AI? You create a new market and it becomes, and you want to become its indispensable foundation. That's how you capture all the economic rewards and the profits. In the case of pharma, you just dominate a market and you defend it. You do not do more. Right. And the key vulnerability is of course architectural shift. So in theory at least an AI market is more competitive. The pharmaceutical market is non competitive even in that theory. Right. So here, this is a slide that looks at strategic lock in approaches. So here basically I'm just explaining how you use actually high entry barriers and you use actually specific lock in effects from patients to patient, ecosystems and other means to create the kind of, same kind of sticky ecosystem that you see in the case of artificial intelligence. Right? And both ecosystems, firms are constantly in a struggle in both AI and in pharmaceuticals, to create ecosystems that become more valuable than the core product itself, because that's the Only way you're going to create powerful value retention mechanisms. Now I'm going to look in the rest of my presentation how pharmaceutical firms actually do it. Right. So because in the pharma sector, first generation, I'm going to look at insulin and small molecules to show you how pharmaceutical firms engineer dominance. My first case is, okay, in the paper, I look at a series of cases to look at these different generations, right? So my first case is diabetes and insulin. So today, if you take global diabetes distribution by region, this is how it looks. There's 600 million diabetic patients globally, roughly 240 million undiagnosed. So close to a billion actually, if you take them into account, diabetes already cost at least US$1 trillion in health care over the last 17 years as healthcare expenditure globally. And a person is dying from lack of insulin globally every five seconds. But so now how did. Actually there are three firms that engineered market dominance in the case of insulin. They started in the late 1990s when we switched from human insulin to analog insulin. The three firms are Novo Nordisk, Eli Lilly and Sanofi. And they actually lock stepped their pricing mechanism from July 1996 to 2016, in the way that is shown in the chart. They did the same thing with the devices. This is how the prices went up practically with the devices. Okay, now this is actually even incomplete. If you take that slide and construct it until 2025, you have continuous innovation in the device area, which is then actually combined with the product. So these three players, they practically captured about 97% of the global diabetes market for about 20 years. They then coordinated dominance with devices from then until now. And how did the winner takes all effect come to be true in this case? So they used the patent mode. So the original patent was expiring sometime in 2016. But if you see here, what happened is that there are patents which were taken by each of these companies that extend actually protection until 2031. Then they also use the device and different kinds of patents around actually the device. So there's for instance, the second table here has three examples. There is a drive sleeve and spring mechanism for dose delivery. End of dose, click is patented. Acoustic sensor for automatic dose is patented. So this is how they actually have extended the patent monopoly from 2016 until about 2040 for those three products. Now there's one slide missing here. Unless it's in a different place. Yeah, here. So what that means for competition. So today, actually in the last few years, the insulin market has changed. There are a couple of new companies that have come in. There's companies that have introduced biogenerics for analog insulin. So that has made the market structure slightly more equitable. And the three large companies have lost a little bit more of their market dominance. They own still about 90% of the global market, not 97. And there's price competition, but not as much as expected. And this is still not going to even out for quite a while because of all the other patterns we saw. Now I'm going to look at a different kind of product, which is monoclonal antibodies, and I'm going to look at how pharmaceutical pharma engineer dominance here. Humira. Humira is actually a drug for arthritis. It's a prescription drug. And in the case of Humira, for instance, you have a bottom layer which is the original patent, which also expired in 2016. You have formulations, which is basically how you formulate the drug concentration buffer stabilizers that go on until 2028. You have method of use patents, which is basically what we did in the case of Humira is they took separate patents for each of the diseases. So they took Crohn's disease, psoriasis, etc, etc, for so many diseases to create patent tickets. And then there are basically patents on injector mechanics, design and user experience. This extends the patents on Humira from 2016 to something like 2036. So 20 years of additional dominance. Now, Humira then became the industry standard because after Humira managed to do that, you see a lot of the drug companies do that for all sorts of small molecules. So here you see Keytruda from Merck for oncology. You see the same thing for different kinds of cancer. Actually, this list can go on and on. In the paper there's a very long list, but I've decided to make it shorter for the 40 minutes that Ken has allotted me. So semaglutide follows a similar path. Right? Semaglutide is another small molecule where right now actually a couple of patents should be going off patent in several countries. It should achieve a generic status, but it's not going to entirely achieve a generic status because you have actually formulation and stabilization patents, you have device delivery patents, you have method of use patents. So that's going to extend the semaglutide monopoly globally to at least 2035, if not more, even though it goes off patent in a number of countries. And that will limit actually the availability and accessibility of those drugs to people who really need them. Now, in the case of monoclonal antibodies, this is a table that shows who invested in the production of. Hold on, I have a problem with my slides. These are not in the right order, so I'll just go on Anyway. So in the case of monoclonal antibodies, you have development timeline, and this slide shows that. So what has happened in the case of monoclonal antibodies is that you had enormous amount of public research investment into creating these products, right? So, and that public investment came here in the UK and it happened in the US and after that you had privatization and patenting, and then you had several early failures. The blockbuster success came much later, actually, when you compare it to the initial discovery of monoclonal antibodies. This is a table that shows you the public versus private monoclonal antibodies investment. So you see, for instance, that in the public sector you had much less investment, but it was actually foundational. Without that kind of foundational investment, you would not have had actually the discovery for product development. And on top of that, the private sector invested a substantive amount. And there is obviously a huge disparity between the public sector investment and private sector investments into this drug, into these kinds of drugs, which I'll come back to in a minute. Now, next generation of drugs are personalized therapies or treatments. Here I'm going to at cart, which is a T cell therapy for cancer treatment. Now what happened in the case of CART is that you have actually the public sector, which again set the foundation. So you had the National Institute of Health in the US which provided the seed funding for the therapy. You had early work on synthetic T cell receptors in the 1990s. You also had early stage clinical trials, which, which was done by the public sector. Then you had private sector commercialization with biotech startups and big pharma acquisition. Right? Now, when compared to public sector, here again the investment is massive in the context of the private sector. If you compare that once again here you see that the public sector played a key role, but the private sector is the one that invested tens of billions of dollars for development acquisition. The CARD strategy is very interesting. When the firms invested the money for product development, they actually moved the dominance game one step forward. So what you see is, you see the core foundation, you see the intelligence layering, you see manufacturing and logistics that are very important, which are already important in the case of monoclonal antibodies. But here you see something additional. You see the clinical ecosystem. So without the data, the clinical data, you're not going to be able to replicate it. So unlike small molecules, card dominance is built on operational excellence and clinical integration. So it creates even more durable modes in the pharmaceutical sector than the ones that we had in the other generations. And the last generation that I actually use in the paper to confer is pharmaceutical platforms. And here a very interesting case study is MRNA. So during COVID 19 we saw certain key differences in the MRNA investments than what we saw for cart, then what we see for CRISPR, then what we saw for monoclonal antibodies, and something else before that. The big difference is that the private sector actually played an important role, but the public sector played a massively important role. So the speed and scale of investment came from operation warp speed in the US and advanced purchase agreements, which guaranteed market demand for these vaccines. And parallel manufacturing was developed as clinical trials were taking place. So we really had speedy entry of vaccines in the market. So the whole manufacturing process was dearest by public sector advanced purchase commitments. And the MRNA platform technology, which was lying on the shelves for decades, was made mature. Now, in this case, what happened was that the financial scale of public sector investment was, if you take the whole of the COVID 19 vaccine development was up almost $12.5 billion right now. This is very, very different when you compare it to cart, when you compare it to CRISPR monoclonal antibodies. And the kind of investment is also important. It's not just the amount of investment. In the case of MRNA vaccines, the investment didn't go into early stage discovery, it went into all the different stages until product development. And this made it very different. Right now, what you have now in MRNA vaccines is also a similar kind of vaccine dominance. So if you really go into the different firms that control the patents and what they control and how they control it, you see actually that you have a core foundation on the MRNA sequences. But the lipid nanoparticles delivery and the next generation LNPs, they become very important. Of course, manufacturing and logistics is very important, and the clinical ecosystem is also very important. But the big difference is, and it actually came out in a paper that Ken produced, published recently with some colleagues, is that in the case of the COVID 19 vaccine, you see increased collaboration, something that you don't see in the case of other previous discoveries. And that increased collaboration comes from the fact that there was public R and D smash throughout the spectrum instead of just in the beginning. So why do we observe this? There are a couple of reasons why we might be observing this, and I'm going to just sort of like go over them very quickly. The first reason is, of course, the model of innovation has changed. So there's been declining public sector RD over the last 30 years globally. And in the paper I compare US, UK and Germany as three countries with significant public sector R and D investments. Here is the German public sector r&d vs private sector private R and D. Okay. And you see a decline actually from 1.2% to 0.8% of GDP over the timeline. But here I compare USA to Germany. USA is the blue line. And. And you see that that's been a much more radical decrease. Right. Okay. And that has fallen down substantially. Now UK started out much lower. The public sector investment into R and D was around 0.65% in the 1990s and it's gone down to 0.35% by the end of the 2010s. Right. So now one thing that of course I'm comparing public R and D as a whole here, what I should be doing is comparing just pharmaceutical sector R and D. So that's in the works. I'm trying to get it done, but it's not so easy because I need to compile data from National Institute of Health, National Science Foundation, Defense and so on. And then we need to see it over time. But hopefully I'll be able to do that in the paper because this is an important point. Another thing is, of course I should also be looking at absolute R and D investments because one can always argue that we are now in technological domains where end stage development is very, very expensive. So it's not enough to compare it and that could skew the scale. So that's also something I should be doing. There are two other factors that explain. One is geopolitical dominance. Dominance. So what has happened over the last 20 years especially is that some companies have become very dominant players in key sectors and economic growth of countries depends on that. And economic superiority of countries depends on that. And because of that there is actually a tendency to protect those gains. Like for instance, during COVID 19, we saw that Biontech or Moderna didn't want to license their technologies. And that was something that nobody contested and the governments of these countries more or less supported it. The other factor is regulatory capture. So pharmaceutical firms are able to capture the law. Because the discussion in pharmaceutical, pharmaceutical industries has always been innovation and R and D versus access. And access is seen as a distributional issue. It's not seen as an issue which needs to be counted in when we construct innovation and R and D models. And that division continues until today. And that regulatory capture between firms that innovate and they lobby for intellectual property and other gains is actually significant. Now what that does it creates actually incentives for mimicking winner takes off effects in the pharmaceutical sector. And to an extent it might also be happening in the AI sector, which we won't know until and unless we break it down into each of these subsectors and look at it. What that also does is that it actually challenges economic thinking on the topic because ideally in a winner takes all market, the advantage is that you will have continuous innovation as a necessity because firms will continuously creatively destruct each other. Right? So you need to see performance based market leadership and you need to see rapid adaptation with competitive outcomes. But in reality you don't see that. You already don't see that in the case of AI because you see consolidation around a few hyperscalers. You see some niche opportunities, but you see a lot of market exclusivity in the pharmaceutical sector. You even see more, greater problems because you see artificial monopolies with little or no technological breakthroughs for large parts. You see privatization of public research in most cases and intellectual property going to private firms when public sector research was the main precursor of those innovations. And you see very strong frictions between innovation and access. And you see a low understanding of the dynamics of industry and policy, see discussions. So in general, we put radical innovation now in the hands of the private sector, both because we rely on these companies and both because we've had declining private public sector R and D. Now you see what does this do? This can create a value of debt for foundational ideas. So in the next 20 years we might not have a new technology to develop because public sector didn't support it and there's nothing to uptake. You can also see a tendency to view innovation as profits and to always specialize in incremental innovation. LLMs as the sole part to AI, for instance, is what we see today. It's dangerous, right, because it might become the only AI paradigm that we have. Then you see barriers to entry and a lack of open science because firms don't want to share. And you see stuff with competition. And I think I need to end. So some final thoughts. We need a more nuanced debate that explores the gap between.
- [29:56]
D
You can take a little bit more.
- [29:58]
A
Really?
- [29:58]
D
You have until quarter class.
- [30:00]
A
Oh, okay. So, okay. I mean, no one told me this before. Okay, now I'm going to take some more time, but this is my last slide, so I'm going to actually go a little bit more into this. So what does this tell us actually? So, so we need a more nuanced debate. Right? Okay. And we need a nuanced debate that explores the gap between the naysayers and those that are for technology. And we need that both in pharmaceuticals, and we need that both in AI, because AI or all of the digital economy has come up with this technological determinism. We always assume that more technology is good, right? But we don't assume more technology might not be good because is not perhaps the right kind of technology. And I think we need a more nuanced debate on that. We definitely need that in the case of pharmaceuticals because we've had until Covid, the tendency to separate access from R and D and innovation. And after Covid, especially with the debate around vaccine production, what we actually see is that we've now made, made it a debate about R and D and innovation and then production, right? So so long as we have innovations and we are able to license it, it must be enough for production and access. And I think what my work shows and the paper shows is that there's a need for a very nuanced debate even in the pharmaceutical sector, because every product category is different. And these product categories, those way we are moving with R and D and innovation and protection of rents, innovation, rents in these different subcategories, it's going to change the way we are going to be able to access health care, we are going to be able to access medicines. So we really cannot think of the world in the old sense of innovative firms and generic firms. We really have to be break it up and think about it in terms of what is it that we want to regulate in the context of monoclonal antibodies or other kinds of biological molecules? What is it we want to do in the case of certain kinds of technologies and what is it we want to do in the case of pharmaceutical platforms? I think also a very important thing that comes through in my research is the link between pharmaceutical firms and high tech AI firms, because there's a lot of compounding advantages of AI firms than actually sharing data and other kinds of innovations. For pharmaceutical innovation and patient health care, the point on regulatory capture is key. Colleague from Harvard, Lawrence Lessig, he's a professor, he said long Time ago in 2000, actually, code is law. He said that because when you write code, you don't code becomes the law, you know. So his point was that once you have some kind of software code, that's the one that decides what the extent of your privacy is. So it doesn't matter if law recognizes privacy or not. It's the code that's the law. And we need to think about the pharmaceutical system sector in similar terms. We need to really start thinking about what is it that is the law in pharmaceutical sector, what is it that actually really matters? Right? You know, if we say we don't want to have this kind of technology dominating a market, if we say we don't want this kind of intellectual property, then firms can create actually dominance through other means. This is what actually breaking up the pharmaceutical innovation into different subcategories in my research shows. So ip, we thought was the thing that needed to be controlled, but it's not necessarily just IP based. So what kind of regulatory mechanisms do we need actually to think through the kind of regulatory capture that we see in the pharmaceutical sector? The third point that comes up, and it seems important, it also links actually to the Nobel Prize in Economics that went to Professor Aguillon here in LSE last week and to the Nobel Prize winning work of Darren Achomoglu from MIT last year. Is this thinking around technological innovation and the fact that we know so little about it and we need to actually, as economists, think differently about technology innovation. Acemoglu, for instance, talks about directed technological change, right? So Acemoglu's point is that technological change needs to be directed in a certain direction. So because firms innovate, but they don't necessarily innovate in a direction that you want them to go, they might innovate in any direction. So a good example is that firms always innovate to cut down on some cost of production. So if you have a firm that wants to cut down actually its production cost in terms of energy, it might want to find a different solution, but it might not necessarily want to invest in green energy. So if a firm says, I'm going to cut down my energy cost is different from I'm going to cut down my energy cost to create, to have green energy, right? That is directed technological change. So we need to think about directed technological change in a different way, particularly in winner takes all markets. Because in winner takes all markets, it's actually the firm that sets the direction of the technological change. It's not society, it's not policymakers, it's not consumers, it's the firm that decides, oh, here's the next big winning thing. I'm going to get a lot of money by doing this, so let's move in that direction. And I think that that kind of thinking, how do we break that and what do we do for that? I think that's important. And to link that to Agyon and Howitz work, and I think that's important because in their 1992 paper which actually won them the Nobel Prize, they talk about the kind of innovation we need. So their point is that innovation needs to be vertical, as in there needs to be a clear step, a move towards the next frontier. And that's what really matters. And when you have that kind of innovation, you make firms obsolete because the firm which is actually in a dominant position is going to be outpaced by a competitor who comes and does that. Now, in the kinds of markets that I've talked about, you don't have that because what you have is that you practically have a firm which has captured a large market share and it does not want to innovate because it wants to keep that market share. And so that's the kind of innovation we don't want as a society because it's not really taking us forward, it's incremental innovation. And then the last point from my paper that I think deserves a lot of attention, which I'm thinking about, but I don't really know in which direction my thoughts are going specifically at this point is that industrial catch up requires everything. We talk a lot about industrial policy and we talk about how countries need industrial policy and how they need to actually really enact different kinds of policies to build sectors. Right now what is happening in one country here in the uk, in the US or in a specific African country most likely doesn't have to do with that country at all, considering these kind of sectoral dynamics, because they are international, they are coming from the outside. So the kind of industrial policy that we've been thinking about is insufficient actually to capture the kinds of markets that we want to regulate and the kinds of negative effects that we want to not have on firms, on consumers and on economic growth. So I think industrial catch up requires a serious rethink in terms of, you know, how do we conceptualize industrial policy? What parts of industrial policy are global and what parts are national? And can we actually break this, you know, vicious cycle by investing a lot in public sector R and D? So those are some of my thoughts and I'll leave you with them and answer your your questions. Thank you very much.
- [38:37]
D
Where do I turn it on?
- [38:42]
C
Mine's on.
- [38:44]
D
Can everybody hear me? I think it is working. Okay, so thank you so much for this very detailed and really rich talk. I think that you laid out the kind of anti competitive dynamics of AI really, really expertly at the beginning and then also show the parallels with respect to the pharmaceutical sector. And I think what I've kind of really took away from the slides and the talk was kind of thinking about a kind of material and infrastructural approach to the economy, right? That firms are not just kind of competing in terms of their products in a market, but they're trying to build a kind of ecosystem, both in a kind of infrastructural way, but in a regulatory way to kind of assert their dominance and to kind of limit entry of other firms into a market. So I think that this is kind of a really important way of thinking about markets and economies, kind of thinking about the broader infrastructure, the broader, broader ecosystem within which development and competition takes place. And I think it's true in other sectors of the economy. If we think about compliance, if we think about the kind of all the private governance within agriculture, being able to get access to a market or not requires all sorts of compliance. And often there are private actors who are lobbying hard of the parameters of that kind of compliance and access. So I think it's a very important way of kind of understanding the way that technology is changing the global economy, changing global value chains and changing the position of different countries and firms within countries. Within this context, I think more broadly now we see this kind of geopolitical rivalry when it comes to technology and the infrastructure of the Internet between the US and China. We're much more aware of how the technology, we shouldn't take it for granted, right, that the standards that we that govern technology and govern data are kind of embedded very much in that infrastructure and have kind of important ramifications about lock in and compliance. So I really, really loved this aspect of your work and I'm very much looking forward to the paper as well. I also really liked sort of one of your points at the end about sort of looking at these kinds of dynamics in a case by case basis, right? Not thinking in kind of big broad brush ways of thinking about how do we regulate AI or how do we regulate the pharmaceutical sector, but really paying attention to how almost every drug has a kind of very different, different context. And we have to think in a specific way about how to, how to do regulation, how to do industrial policy. And I think this is so important. You know, I also have found this very much with thinking about data for development or AI that we kind of have these big broad brush claims about what AI is doing to the economy or what data is doing to the economy. But actually there's a lot of friction that there's a lot of specific sector things going on and we really need A kind of varieties of digital capitalism or varieties of AI approach, I think, to understand how things vary across space. So you've given us interesting things to think about. I did also have some questions for you. So my first question is, kind of a big question that all of the other questions go group inside of, is I wanted you to kind of take a step back and flesh out what you think are the most implications for all of these developments in terms of development, because we are a development studies department. So what's at stake for developing countries and low and middle income countries in these kinds of changes in the R and D environment and the regulatory environment? And my second question is also specifically as the CEO of the African Pharmaceutical Technology foundation, where do African scientists, universities, researchers, consumers and companies fit within these changing dynamics? Is it the case that we see from donors a kind of push for policy making and regulatory compliance solely focused on access and kind of safety and costs and availability, or are there kind of some commercial implications also relevant and salient? If we think about the kinds of potential jobs, the kinds of creation of companies, and do we see variation across Africa in terms of how much domestic actors can really kind of integrate themselves in a kind of commercially strategic way within these ecosystems? My third question was to expand a little bit more and maybe to take one of the specific examples in your paper, like how can the public claim greater control over the direction of research? I can understand where you have a case where there's been a lot of public investment. We think about Mariano Matsucato's work about, you know, if you're paying, if the public is paying for something, it gives them some kind of leverage over the use of that technology. But in some of the cases you presented that where there's huge private investment and not a lot of public investment, how does the public reassert control? What are other particular kind of choke points or lever points that are still at our disposal? And I presume the kind of relative power of different governments varies enormously. Maybe the US government has quite a lot of power. Maybe the Kenyan or the Rwandan government has less power. The fourth question is around data within the pharmaceutical sector. So we were kind of talking before the talk about how I often get frustrated that there's a kind of discourse about how data is being centralized across the, across the board through these platforms and technologies. But actually when it comes to specific sectors, we have a lot more fragmented data in terms of the kind of proprietary interests around data. And if we think about in the health sector, all sorts of kind of privacy Regulation, national boundaries in terms of data, maybe hospital boundaries even. So I'm kind of curious to, to know within this kind of health sector or in the pharmaceutical sector more specifically, do we see the emergence of kind of data centralization within companies that are already established within the sector and kind of technology companies from outside are struggling to get access to that data? Do we find kind of new digital pharmaceutical companies that have some kind of cell sector specific focus but is still not kind of a big technology company? Or do we have big technology companies able to get access to that data and develop kind of sector specific algorithms and technology around that? And then I feel like I'm asking too many questions. But I mean, the last question was around training and skills. At the end of your presentation you mentioned this kind of hype is pulling talent into this sector and to the maybe disadvantage of other sectors. So how, you know, have you seen evidence of that within the kinds of networks that you're in, particularly in African countries, that you see talent going in specific directions rather than others and whether there's any, anything we can do to kind of redistribute talent in a more kind of public spirited way? So I don't know if you want to answer some of those. You can choose the ones that you feel are most interesting to you.
- [46:52]
C
Why don't you do a couple just quickly and then we'll open it up because we have a nice audience here and there's roughly 100,000 people online. And so we'll.
- [47:02]
A
Okay, okay, so thanks Laura. Yeah, very quickly. So what does this mean for development? Yeah, so I've said what it means, I think for industrial policy and industrial catch up. Right. So for me, industrialization is the most essential thing for development. So if we don't industrialize, we do not generate economic prosperity for countries. And in a context where a number of low income countries have not been able to industrialize at all over the last, at least three decades, if not more. The current situation is actually really, if you see sectors such as this, it's really bad because it means also that you don't have basic health security if you cannot actually produce your own drugs and if you cannot have access to health care. It also means that you do not have information security, right. If you do not have any role to play in the digital economy, because AI is actually the one that's all pervasive and it's controlled elsewhere. And so I think that this has drastic implications for the way we think about development. And the irony for me is that at a time when these countries kinds of topics should form part of international discussions, we seem to be sort of like, you know, moving towards a more disintegrated way of thinking about international cooperation. Most international agencies today are thinking about what to do, how to survive and deal with the budget cuts. So I think that it has massive implications for development and some something that I think that we need to work to unpack. How does Africa fit? This is an interesting question. I want to give a case study of what happened in Africa when the two companies which had MRNA vaccines did not actually agree to license. So the WHO decided to set up MRNA hub in Africa with the help of a number of donors in order to make it possible for African countries to produce their own MRNA vaccines. At first the WHO announced the MRNA hub with I think an 18 or 20 month deadline to product development and production. But after the hub was announced, both Biontech and Moderna refused to license this, this technology. So when there was no company ready to license the technology, you had to practically get together a pool of universities in South Africa which reverse engineer Moderna's COVID 19 sequence. I think I forget what was called, but they reverse engineered the key sequence and they actually transferred the technology for product development after reverse and sharing that to a private biotech company based in Cape Town. So this actually extended the life of this MRNA hub from 18 to 20 months to 5 years or something. In the meanwhile, Covid has come and gone. The MRNA hub exists and there's been a massive amount of public R and D that's gone into it and a lot of skills that are being built there. And I think that's a very good model for us to think about, really to think about decentralizing public R and D and creating notes in the developing world on how actually we can create the same kind of economy effects that big firms are able to create in the West.
- [50:45]
D
Was that made possible because it was a health emergency though?
- [50:49]
A
Yeah, yeah. And it would not have come through if not for large investments from big countries and philanthropic foundations. So which is again something we need to think about, what's going to happen if we don't have those investments.
- [51:04]
C
Fantastic. Well, thank you very much, Pad Ranchi for your talk and Laura for your comments and questions. And why don't we open it up to the floor? If people have any questions or comments, just raise your hand and there's a roaming microphone.
- [51:20]
A
Please. Thank you very much indeed. I'm a former law enforcement intelligence analyst. I always find events like this hugely informative because of course it's finding things out. My question follows on from my work. Were there any things that particularly surprised you in your findings or conversely that did not that you had a hunch you would come across certain patterns and you found them or things. Things were surprises. Thank you. We're all. I say that because we're all geniuses with hindsight, don't we?
- [52:05]
C
Thank you.
- [52:08]
A
Sure.
- [52:08]
C
Well, people are coming up with more and then.
- [52:10]
A
Yeah, okay. So actually, you know, in the case of, the reason why I went so deep into the pharmaceutical sector to break it down was that people always talk about the winner takes all dynamic and they believe it comes purely from intellectual property. And I had a hunch that it's not possible that across all product categories in the pharmaceutical sector, you are having those effects just because of intellectual property protection. So I decided to go, you know, somehow, you know, therapy versus devices versus products versus platforms to check it out. So in a way I had a hunch, but it sort of helped me to clarify my thoughts. And, and I think that the sort of inter industry analysis sort of makes a stronger case for thinking about it from a perspective of how do we allocate our resources properly into innovation and R and D rather than thinking about innovation and R and D is very important. We're already doing enough to support it and we don't have to think about access because that's going to flow automatically. So it should have shifts the debate to hopefully a more objective one. Great.
- [53:34]
C
I'm going to question an online question and then we'll come back to the audience. This is from Nuno Nunes who asks what can be the role of international organizations to democratize innovation? As someone who works for an international organization.
- [53:52]
A
Yeah, actually I'm a lot of. Well, I'm not very optimistic about this because I've worked for a very long time in different capacities in international organizations and practically the mandates of a large number of international organizations are set in a way that you don't influence innovation. Because like I'll give an example, when we set the MDGs agenda, the innovation was a subpart of Goal 8, but really you were to focus on reducing child mortality, maternal mortality, access to sanitation, etc, etc, but not to improve innovation outcomes. But the same thing with the number of international agencies and their mandates. So what? What can they do? I think we should really sort of try to coalesce around thinking about how to make firms more innovative. And I think if we start thinking around that, we might do things which are also useful from a developmental perspective. I think these two things are not necessarily at loggerheads with one another as we think they are.
- [55:06]
C
Thank you. Thanks.
- [55:11]
B
Really interesting talk. In your research, have you thought of any sectors or even examples within the pharmaceutical sector where governments and regulators have been able to effectively overcome this market dominance and bring access benefits? I guess you talked a lot about COVID vaccines. I'm thinking of AstraZeneca actually had comparatively quite a good access policy. But then I think some of their performance as a company did quite poorly as a result. But I don't know if there are other examples, for example from energy sector and solar that you'd thought of. Thanks.
- [55:50]
C
Why don't we collect a couple? Okay. I know. Kate, you had.
- [55:54]
B
Yeah, thank you. I just wonder whether.
- [55:57]
C
Yeah, go on. And then.
- [55:58]
B
Yeah, sorry. I just wonder whether you think this winner takes phenomenalist kind of inevitable consequences. Government picking winner rather than picking picking missions. So thus basically government intentionally build up and protect the so called winner or perceived winners in certain industry to get quick win and the visible win in certain developments certain industry. And I wonder whether there is any. Do you have any view on alternative approach to technological innovation that focus or based on let's say open protocol, open in. In the case of AI that will be open source AI model this kind of innovation. Thank you.
- [56:53]
C
Thanks, Kate.
- [56:55]
E
Thank you for a really interesting talk. Excuse me. I have a question about regulatory capture which often involves the people of these private sector firms embedding themselves. Sorry, embedding themselves into the regulatory committees and bodies in which these decisions are made. So I'm wondering how much scope you see for addressing regulatory capture when many of the people who are part of the problem of regulatory capture are in the public regulatory bodies themselves and even in some of the larger international organizations. I thought the case of AstraZeneca is quite an interesting one, where a public needs the developed vaccine was required under WHO pressure, led by a leading private sector actor to link up with a private sector firm before they could participate in vaccine development. And just another quick question about the use of proprietary devices to extend patent control. To what extent can delivery just go back to other systems? So if they have new methods for delivering insulin once the patent runs out, is it not possible for people to say, I'm going to use hypodermic needles and make a generic one. I don't care if you have a proprietary device. We're not going to use that. We want to use the insulin whose patent has expired. Remembering of course that insulin itself was developed and sold into the public system for $1 to the University of Toronto. And there's absolutely no reason why it should be now.
- [58:45]
A
Hi, I'm interrupting this event to tell you about another awesome LSE podcast that we think you'd enjoy. LSE IQ asks social scientists and other experts to answer one intelligent question like why do people believe in conspiracy theories? Or can we afford the super rich? Come check us out. Just search for lseiq wherever you get your podcasts. Now back to the event.
- [59:13]
C
Thanks, Klabashi. Why don't you hit those three and then we'll go for another round here?
- [59:18]
A
Okay, so your first question was you asked me about AstraZeneca. Right. And you wanted to know if. If I knew similar examples from my work. Right. So I think AstraZeneca is a really fascinating example because if you trace the history of AstraZeneca, how it started, what it does, and where it is today, you also see the impact of financialization on the way AstraZeneca approaches access. Right. So AstraZeneca was way more invested into a lot of different sort of public health based disease categories previously than it is today. And one of the reasons for that is, of course, financialization of the pharmaceutical industry. And this increasing pressure to deliver on shareholder returns. And what is good for the shareholder is not necessarily good for innovation. It's actually more important nowadays for pharmaceutical companies to retain incumbent advantages, which I think explains a lot of this winner takes all strategies that they come up with. So I do see a lot of companies that are sort of like fundamentally sympathetic towards access goals, because I don't think it's useful to paint a picture that all big pharma companies are bad. They're not. Everybody is trying to operate in an ecosystem which is structured in a certain way. And a number of companies are thinking seriously about access. Sanofi has done a lot of tech transfer for different vaccine categories to Africa and Asia. You've got a couple of other companies that really have very serious access campaigns. But on the whole, my point is, and the research points to, as Laura said, that it's actually the ecosystem in which we operate in and the infrastructure around which this kind of innovation thinking is built, which is the one that's very problematic, actually. So that's about that. Regarding picking winners, right? Yeah. States, countries always pick winners, right? Industrial policy has tended to a large extent to pick winners. And that can distort the market. Because when you pick winners, you're sort of like setting the stage for some companies to win at the expense of others. But at the same time, you know, the winner takes all is different from picking winners because the winner takes all is something that happens on its own. So when you pick winners, you're not going to, you're going to create a certain environment. So take for instance South Korea's industrial strategy in the 70s, 80s and 90s, right? So when you pick winners, you're going to create a certain sort of like an overall ecosystem for innovation to flourish and you're going to create a carrot and stick approach for firms. Right? Okay. But here what's happening in winner takes all markets is that you're going to have one firm that comes in and because it came in first, it's going to just swoop away all the advantages. The rest of the firms which have invested massively, they're not going to be able to get the gains. I'll give you an example. During COVID 19, there were many more companies that were investing in MRNA vaccines. Sanofi was one of them. But the moment that Biontech and Pfizer announced their partnership, Sanofi announced that it will no longer pursue the R and D on the MRNA vaccine. Why? Because it's a winner takes all market. Every company knew that the first company that's going to come in with an MRNA vaccine is going to get the global advantage, right? And once you have that dynamic, you're going to see other firms not invest. And that's the situation you don't want to have. Because that's not good for society, that's not good for, for innovation. Regarding your question, yes, that's a very big problem because regulation has been co opted by people who lobby and pharmaceutical company companies have very strong industrial associations and they have very strong lobbying, not only with regulators, but also with patient care groups. For instance, in the case of insulin, I did that work over several years because over and over again a lot of different people came and asked me to study the global insulin market. Because we were going to reach the hundred year mark of insulin discovery and insulin was not available to most people who needed it globally. And what I found is that for instance, in the US there's such a big short shortage of insulin for people who need it. And what I found was that it was not just regulators that were captured, it's also actually doctors, patient care groups, everyone's captured into one kind of product or mechanism. In the insulin sector, what is very interesting is that. So the device market is very important because the global medical devices market is not as competitive as we think, because it's very hard to produce precise medical devices. So when you have medical devices that sort of like are very good, the competition for that is less. That's because to produce medical devices you need completely different skills than, than what you need to produce a drug. Right. Intuitive. But those kinds of skills are very hard to come by. They're very hard actually for companies to come by in a large number of countries, particularly in the developing world. So if you actually take a product, put it together with a device and you patent it, and then you create actually a patient ecosystem around it, it's almost impossible to break that. For a new firm that comes in, there were no old devices. These devices came with analog insulin. So the market started to become like that when we switched from human to analog insulin. Analog insulin has certain advantages over human insulin, especially when you have type 2 diabetes. And that is actually very useful, actually to administer an alternative analog insulin. But what we have now as a global situation is that because of the shift towards analog insulin and this massive commercialization drive from these companies, right, there are lots of people dying in the developing world whose lives could have been saved on human insulin. So it's like if I was going to die of insulin and if you told me I can give you perfect analog insulin, but you can't access it, but I'll give you human insulin and you're going to destroy, struggle for a few years, I would definitely choose the human insulin, but it's not available. So that's, that's the situation we got to.
- [66:25]
C
In the middle here.
- [66:31]
E
Thank you for your presentation. You talked about how leading companies in.
- [66:36]
A
The AI and pharma industry are the.
- [66:38]
E
Ones that kind of set the direction and not the policy, policymakers, the public and governments.
- [66:45]
A
But so I just want to know.
- [66:47]
E
Like, what is your opinion on what.
- [66:50]
A
The role of public policy should be in these industries and how we can.
- [66:56]
E
Kind of create policies that don't stifle.
- [66:58]
A
Innovation but still protect people from the.
- [67:01]
E
Negative effects of the industries.
- [67:06]
B
Hello. Thank you for your presentation. Just to add on to that question, specifically about patents, I just wanted to know your thoughts about the role of governments like the UK and usa, where we have consumers who presumably are contributing quite a lot to these, for example, pharmaceutical companies in terms of the affluent customer base in these countries. And so does that give us more power, more to regulate against these patents and promote innovation in those industries comparatively to the developing world? And then, and then, if so, are there parallels in terms of how we can go about regulating AI? In the same vein.
- [67:51]
A
Thank you again for the talk. Maybe just to add on to the first question, if you have a specific example, maybe specific to, particularly in the, in Europe where you think a policy has tried to address specifically this kind of public, private innovation collaboration for drugs, if you have an example of a policy that you think actually has made a good step in the right direction.
- [68:16]
C
Back to you, Padmashree.
- [68:18]
A
All right, so to your question, what should we be doing in terms of public policy? Right. So I think what we need to be doing is I think we need to be doing way more sort of like we need to understand the industry we're trying to regulate, which is something mostly we don't do in public policy. We try to just regulate. Right. And I think that it's not just my work, actually. There's a of lot, lot of really interesting work and I'm happy to share the links or the authors with you. There's a whole new set of insights coming in which talk about really looking at the industry that we're trying to regulate. And I think having that informed discussion is very important. It's also important for AI, for instance, because most of the stuff that we talk about AI is really like, should we regulate AI, should we not regulate? You know, I mean, that is just such a simplistic way of thinking about regulation. You know, what should we be regulating in AI and what should we not be regulating and where should we be focusing is probably something that we should be thinking about. The second thing is that law always, or public policy comes as an afterthought. So something's going on and then public policy comes much after that. And we saw sort of play catch up with what's going on in every technological domain with public policy. And I think that going hand in hand is really important. Right. Which talks about more interaction, more collaboration between ourselves as scholars, thinkers, experts, but also with the scientific community. So those are like two things that I think we should be doing. I mean, there are a couple of, of questions that speak indirectly to something that I mentioned in the paper, but I think we, we didn't have time to discuss. I think that the model of innovation is very important that we opt as a society. I think this is something we can already anticipate and we can foresee. So for instance, we can't do everything that private firms do through public investment. That would be stupid because it's, it's sorry for you. That would not be really useful because we can't do that. Private firms have a particular value for innovation and we need to maintain that. The question that we need to pose ourselves is that how do we make sure that we create the right environment for private firms to uptake public discoveries and to create products, but in a sort of a competitive environment? And I think that to get the right mix of public versus private is where we are sort of struggling and that's where public policy probably needs to focus on. Your question was on. Sorry, I forgot your question. Could you please repeat it? Yeah, so. Right. Yeah. So we have a tendency in the world of intellectual property, particularly in place, the first pharmaceutical sector, there's a lot of IP that gets sort of like first granted to firms in the US and then a lot of the other firms, they then go and approach the other countries and then you sort of like sort of evaluate the patents. And I think that looking at patents more stringently will be useful. Looking at novelty and inventive step when firms apply for patents is going to. To be useful. But this is this very contested area in the pharmaceutical innovation sphere. So if you reopen the question of we need to look at novelty and inventive step, a lot of people will not like it. But I think at the same time, what might be useful is also going beyond pharmaceuticals to AI is to think about what is the kind of society we want to have, what's the kind of innovation we want to have. I think that's very relevant, right? Because for instance, sense whatever we do in the pharmaceutical sector, if AI grows as it's growing, we're going to have unemployment, we're going to have a lot of unemployment in different, to different extents, in different kinds of things that we do in other sectors. Right? Okay. So as a society, we need to be able to think about it. In the US discussion is different because the US has and houses a number of these companies which have certain global advantages. Should the discussion not be nuanced in other countries in Europe? Should it not be nuanced in other countries which don't want to lose employment because of new technologies? I think that's the kind of sort of decentralized public policy discussion we need to have. Then there was the third question about whether I have actually examples about good public private collaboration. Right. In Europe. So I think that there is recently some countries and in some cases we're looking at. So what happened predominantly during COVID 19, but it has happened before, but it got legitimized during COVID 19, is that we don't question anymore when private firms take on public research and patent it. Right. And that started with the whole MRNA thing because a lot of private companies also viral vector Vaccines and so on. They took public research, they produced products and then they patented it. Now what I see in a number of cases is that there is now access provisions linked to doing this kind of stuff. So if you get patents on something that got discovered using public funding, then you need to ensure that there's greater access to. So there are conditionalities. Right. What we could do is also make it better. It's not about access. Right. You, you need to sort of collaborate or you need to license. We can add things and I think that would be good, sort of like precedent for us to start thinking about other innovation models. Because otherwise you're giving public taxpayer money and you're going to create proprietary innovations for the private sector.
- [74:23]
C
Patma Rashi, there's a question question online that actually builds off this third question here that you just answered. But this person, Isidin, who's a UCL graduate from Jakarta, asked the question essentially the same question about the role of the government, but specifically in developing countries. So rather than just some examples from Europe, are there examples or suggestions that you have for how developing countries, governments can do this?
- [74:48]
A
Sure, yeah. So I think that. So in developing countries, is that specific to the pharmaceutical sector? Yeah, I can answer for the pharmaceutical sector. So in developing countries, what we have been doing until now is we've been thinking about innovation in terms of generics, right? Because we don't have a possibility to enter the innovators market so easily and it's getting more and more complicated with, with the new technologies and platforms. So a lot of the thinking is just that we can have a couple of firms, they produce generic molecules and you know, that's going to be useful for us point of view, health security, or that's going to be useful for us in terms of having economic development with the pharmaceutical sector. I think both of those kinds of assumptions need to be rethought in the current context, because generic firms. For instance, my work on the Indian pharmaceutical sector for the last 25 years or so shows that, you know, Indian firms are probably the most competitive generic pharmaceutical firms in the world today. Perhaps the Chinese exist too, and there are some, but as a sector they are competitive. And even those firms which capture a huge share of the volumes market globally for generics are unable to get the R and D investment needed to do clinical trials for new vaccine introductions. Right. Okay. Now this is the scale of investments we're talking about and they're not able to get it. So we need to start thinking differently about what is the kind of industry that we want to build in the developing world, and we need to think massively about how the state can support it, where to bring in that kind of public funding. Like, I mean, I don't know, take quantum computing today. I mean, Western countries are pouring money into it. Billions and billions of dollars. I mean, you take the developing world, there's no real, sizable public investment into these things. And I think that we need to start rethinking them.
- [77:01]
E
Thank you so much for the talk. I was wondering. So do you think that there is going to be an opposition against these kinds of policy that you mentioned? So, for example, what you just mentioned about, like, the promotion of innovation and where or who do you think this will come from and how do you think you should overcome it?
- [77:17]
A
I didn't get the second part of your question acoustically.
- [77:21]
E
So basically, where do you think that this opposition might come from and how do you think we should overcome this?
- [77:27]
A
So, yeah, there, there is a lot of opposition to that kind of thinking. The kind that I'm suggesting, it comes from industrial associations. It comes from. It might come from countries themselves who have dominant companies. Right. And it comes also from regulators. And it can also come from people who, who don't understand the sectors. Right. Because if you don't understand the sectors, you also don't understand the need to look at it in that way. But I think that there is a value to thinking differently about it, and there's a value to doing it, because every country needs its own model of development. We can all be doing what one country is doing because it won't make sense, because we're not all at the same level of development. We don't all have the same unemployment profiles. We don't even all have the same sort of sectoral specialization profiles. So we need to be thinking about what's really good for us as a country and then doing it in our own individual countries. Thank you. You just mentioned earlier how a lot of the companies that develop these technology are based in the United States, which it makes sense to think that that would give the United States government the ability to sort of influence how the tech is developed and distributed. And you just mentioned that every country has its own development needs as well. So how can other countries that rely on the technology that's based in the United States assert some control over making sure it protects their own public interests as well? Yeah, that's a difficult question actually, to answer because you have already these companies that have accumulated these advantages. They are, in the case of AI, a lot of them are Based in the US right now. What do you do if you're a late comer? What do you do if you didn't get there first? And that's a question that a lot of the other countries are now struggling with. So if you see, for instance, Germany actually started to massively invest into public R and D over the last few years because of that realization, a lot of European countries are doing that. One way to do it is public research. Of course. The second way to do it is really actually think around more collaborative models. So that's the other thing. The third one is the network advantages are difficult to break. So we need to start thinking about how do we get these companies to share or to. How do we. Actually, there's a lot of cases that happen now in Europe on competition ground against actually data and AI companies, which gives us a good way to think about, okay, should we actually regulate it? Should we, you know, try to break it up elsewhere? How do we actually try to break down these monopolies in a way that they do not stifle innovation in every country?
- [80:15]
C
I think you might have the last question here. Was there someone over there? We can have two last questions. Just person.
- [80:25]
B
Thank you. So in these global winner takes all markets, do you think it's possible for a country's regulatory, regulatory regulation authority to order that a company should be split up like a global company? And if it is possible, would it be worth it? Would losing out on those economies of scale be worth the possible benefit of competition?
- [80:49]
A
The first part of the question was.
- [80:51]
B
In global winner takes all markets, is it possible for a company, a large firm, to be split up by a regulation authority?
- [81:00]
A
By regulation authority, yeah.
- [81:03]
C
And then.
- [81:06]
B
So a company specifically called Open insulin is trying to create an open source genetic code to create analog insulin and then make modular reactors in.
- [81:17]
C
Cities, so then it's widely available.
- [81:19]
B
Is a strategy like that appropriate for things in, in the pharmaceutical industry?
- [81:28]
A
Okay, so first your question, right? So you should. Is it, what is it? Your question is, is it viable? I'm sorry, I couldn't hear it acoustically. To create a new large firm in. To break up a firm. Oh, so your question is, would a country's regulatory authority break up a large firm itself? Can it? Sure, yeah. I mean, if your competition policy allows for it and if it's anti competitive, you can do that. And there are cases actually where this has happened in the past, but it doesn't happen anymore. And I think the reason for that is really this issue of geographical dominance because countries don't want to lose those lead advantages they get from these large firms. So they'd rather not break them up and they'd rather try to regulate it in a different way. But yeah, technically, yes. But here's the dilemma, is that you see globally, if you see a sector, you see one, two or three firms distorting that sector from a global perspective, it might be useful to break up those firms, but the country in which those firms are situated will not be ready to break up those firms because those firms are very important for the gross GDP of that country. So you've got a clear conflict here. So. So you'll have to deal with it in a different way. For your question. Yes, I think that those are really useful, sort of like modular manufacturing facilities or thinking about production in modular ways actually is very useful because it can also help to decentralize certain dynamics that we see actually in these sectors. So if you would have more modular production facilities, model of production model models. Right. You could actually integrate more collaborators in a particular process. You could decentralize different stages of the production. You could create a different kind of value chain. And if you can do that, you could also increase the knowledge below us and the knowledge effects in these sectors, which tend to get concentrated and specialized in certain nodes. And I think that there's this value to it. And, but the only thing is that, and it goes back to the question Laura asked me, which I couldn't answer, is sometimes the skills might not be available for a lot of the people to participate in these notes in different countries or in different contexts. But I think that, but theoretically it's actually a really good idea to sort of decentralize that because it always leads to more democratic outcomes.
- [84:11]
C
So. Pat Marsher, we're going to wrap up in a sec, but I have a question here. There's a lot of students Here who are MSc students at LSE who over the course of their year are going to be working on an MSc dissertation, which is roughly a 10,000 word research paper, and on a topic of their choice, independent study. And so here's this hot topic technology capture. If you were a MSc student at LSU or if you were advising MSc students at LSE, what do you think are the cool topics that you would suggest a student work on that's doable in a roughly eight month period?
- [84:54]
A
Okay, I have a lot. I don't know if. Yeah, so actually I think one really interesting topic that I think we haven't put thought about is about pandemic innovation. Right. Okay. So if you're working on msc, I would say it's a really interesting question. Should we have a different model for pandemic innovation than we have for regular innovation in pharmaceutical R and D? Should we not? Does the same system work? So that's an interesting question that I've been thinking about. Another question that I've been thinking quite a bit about is what should countries do who've not made it to build their pharmaceutical sectors until now? Should they be actually, like, I don't know if you all know, but a lot of countries are now investing in local production, local manufacturing, and so is it worthwhile for these countries to invest? And if they do, how should they invest in, you know, building local production or local manufacturing capacities for the pharmaceutical sector? And then a third question is, which is the kind of model of innovation that we might want to think, which promotes more collaboration, you know, collaborative. What kind of content, collaborative model of innovation can you think in both AI or pharmaceuticals? Because I think that often students like dystopian topics than utopian topics. And I, I want students to think more utopian. It's harder to come up with ideas for a better world. It's easier to believe that everything is going to go wrong. And, you know, so, so yeah, those are sort of great, the topics.
- [86:47]
C
All right, well, you thank. Thanks a million. So thank everybody for your attendance and your participation tonight, and especially those also those who are.
- [86:54]
A
You can subscribe to the LSE Events podcast on your favorite podcast app and help other listeners discover us by leaving a review. Visit lse.ac.ukevents to find out what's on next. We hope you join us at another LSE event soon.