Latest / World Food Prize Podcast: Dialogue That Feeds the World / World Food Prize Podcast with Dr. Michael Kremer, 2019 Nobel Prize Laureate
Transcript
- 0:11SPEAKER_00: You're listening to the World Food Prize Podcast, dialogue that feeds the world.
- 0:16SPEAKER_00: Twice a month, we bring you inspiring conversations with the leaders, innovators, and change makers working to increase food and nutrition security.
- 0:26SPEAKER_00: From groundbreaking research to global initiatives, we explore the ideas and actions driving food and nutrition security, sustainability, and opportunity for all.
- 0:37SPEAKER_00: Join us as we uncover solutions, share stories, and work toward a more food and nutrition secure future.
- 0:46SPEAKER_00: The Wellmark Foundation is proud to present the World Food Prize Podcast as part of our commitment to improving the health of all Iowans and South Dakotans.
- 0:55SPEAKER_00: Access to nutritious food is essential for living a long, healthy life.
- 0:59SPEAKER_00: And this partnership underscores the importance of advancing a food and nutrition secure future for all.
- 1:07SPEAKER_00: Please welcome Michael Kramer and your host, World Food Prize Foundation CEO, Thomas Vilsack.
- 1:14SPEAKER_02: Welcome to the World Food Prize Podcast.
- 1:17SPEAKER_02: Dialogue that feeds the world.
- 1:19SPEAKER_02: Each episode highlights people, ideas, and innovation shaping a more food and nutrition secure future.
- 1:24SPEAKER_02: Access to food and its quality affects global stability, economic opportunity, and human well-being.
- 1:29SPEAKER_02: Today we're joined by Professor Michael Creamer, Director of the Development Innovation Lab at the University of Chicago.
- 1:35SPEAKER_02: He is the joint winner of the 2019 Nobel Memorial Prize in Economic Sciences for his work on experimental approach to reducing global poverty.
- 1:43SPEAKER_02: At the Development Innovation Lab, he develops and scales innovations in agriculture, education, water treatment, and microfinance.
- 1:50SPEAKER_02: These efforts have the potential to reach at least 100 million people in low and middle income countries.
- 1:56SPEAKER_02: His work has led to real world impact.
- 1:59SPEAKER_02: He helped design and advance market commitment that has expanded access to life-saving vaccines in 60 countries.
- 2:06SPEAKER_02: His work has informed national dewarming efforts in India that reached over 275 million children each year.
- 2:21SPEAKER_02: He continues to design and guide practical funding models that help scale proven innovations, including development focused investment funds.
- 2:29SPEAKER_02: Professor Kramer, thank you for joining us today on the podcast.
- 2:34SPEAKER_02: Professor, I cannot thank you enough for taking the time from what I know is an incredibly busy schedule to visit with us uh today.
- 2:42SPEAKER_02: Well, you're doing such important work and it's so expansive that I'm I'm not sure we have enough time to cover it all.
- 2:49SPEAKER_02: But as a way of introducing you to the audience, tell me a little bit about what you do at the university because it sounds like you're you guys are in the problem-solving business.
- 3:00SPEAKER_01: Thanks so much, and it's wonderful to be here.
- 3:02SPEAKER_01: Thanks for thanks for taking the time to talk.
- 3:04SPEAKER_01: I so I'm uh I'm a development economist, so I study the economics of low and middle income countries.
- 3:12SPEAKER_01: And you're right, a lot of what I try to work on is it's actually very practical.
- 3:17SPEAKER_01: Economists have a reputation for sometimes being ivory tower, and I'm a fan of the uh of that side of economics as well, but a lot of what what I work on is is very applied and about concrete problems.
- 3:30SPEAKER_01: And I lead a lab at the University of Chicago called the Development Innovation Lab.
- 3:36SPEAKER_01: And what that lab does is it works with partners, usually governments, but sometimes firms or NGOs, to look at concrete problems that they're facing and that are related to the challenges of global development or global health.
- 3:50SPEAKER_01: And um and we try to understand what issues they're facing, um, and then uh what ideas they have about how to address them.
- 4:02SPEAKER_01: And then we try to work with them, try and draw a global experience and related uh challenges.
- 4:07SPEAKER_01: Then typically the way we we try to help move forward with them is they identify a number of, after this process, they identify a number of ideas that they actually want to try out, and they then try several of them, maybe just one of them, maybe a a suite of them.
- 4:26SPEAKER_01: And we work to help them set up um an evaluation system that's very much like what tech companies do in A B tests or what pharmaceutical companies do in a randomized controlled trial, but adapted to the situation of uh of development economics programs and to whatever their constraints and needs are.
- 4:47SPEAKER_01: And that allows them to learn which approaches seem to be succeeding, but and which ones aren't, but also to learn something about the reasons why.
- 4:56SPEAKER_01: And we don't see this as we see something that comes at the end of the process of you got an A or B, but rather part of a continuous learning process of continuous improvement.
- 5:09SPEAKER_01: So maybe you find something doesn't work, but you understand how you've designed the feedback mechanism so you understand where in the chain of the theory of change this broke down, and maybe you say, oh, we need to tweak that, let's give that a try.
- 5:23SPEAKER_01: And maybe you go through a range of uh of of different alternatives till you've got something that you're comfortable, that the organization's comfortable scaling.
- 5:32SPEAKER_01: And that's ideally not the end of the story either.
- 5:36SPEAKER_01: Then there's further challenges that they may want to take on.
- 5:39SPEAKER_01: So we we hope to create long-term relationships in that way.
- 5:43SPEAKER_01: And uh yeah, that's that's the basic approach we take.
- 5:46SPEAKER_02: You know, one of the partnerships, I think the first time you and I met was in the context of uh then the Aim for Climate uh initiative and and uh the announcing of the Aim for Scale, which was an opportunity for the United States and the United Emirates to work together to encourage a number of countries and innovators and companies and nonprofit organizations to try to address a pretty big issue, climate and food security.
- 6:11SPEAKER_02: You're now still involved, if I'm not mistaken, in the Aim for Scale initiative.
- 6:15SPEAKER_02: Maybe you could uh spend just a minute or two talking about the work that you you're doing in with reference to Aim for Scale, what you're seeing out of that work, and maybe some of the examples of some of the innovations that hold promise.
- 6:28SPEAKER_01: Sure.
- 6:28SPEAKER_01: Well, the way I got involved in that was through something called the Innovation Commission for Climate Change, Food Security and Agriculture.
- 6:37SPEAKER_01: And uh that was supported by a number of donors, but uh the Gates Foundation among them.
- 6:41SPEAKER_01: And I was chair of that of that commission.
- 6:45SPEAKER_01: And what the commission did was we tried to look through a number look at a number of innovations, and we tried to look at the rigorous evidence, look for innovations with rigorous evidence of both of impact.
- 6:57SPEAKER_01: When we say rigorous evidence, that's typically not always, but typically from um from randomized controlled trials or A-B tests from some something that would make it possible to identify the causal impact of the innovation, disentangle it from other factors.
- 7:11SPEAKER_01: So we wanted evidence that the innovation worked, that it was cost effective, but also very importantly, that this is something that there were practical ways to encourage farmers to use and the farmers wanted to use.
- 7:24SPEAKER_01: There might be some seed that performs wonderfully in in a test on a research plot, but if farmers won't adopt it because they don't like the taste or it's hard to market or requires a lot of training to grow, then that's not an innovation that's poised for scale.
- 7:42SPEAKER_01: It may well be worth further research to try to improve, but it's not ready to scale.
- 7:48SPEAKER_01: So we identified a number of such innovations and that were not yet scaling.
- 7:54SPEAKER_01: And I think that's one important finding.
- 7:56SPEAKER_01: There were quite a range of these.
- 7:57SPEAKER_01: How did we connect to Aim for Climate and then Aim for Scale?
- 8:01SPEAKER_01: We were introduced to people in UAE when they were um when they had the presidency of the COP, and they were very excited about trying to find something they could do at COP on food and agriculture and food systems.
- 8:17SPEAKER_01: And they took a look at the list of innovations that that the commission was developing, and they chose seven um that they wanted to uh focus on as uh priority areas.
- 8:29SPEAKER_01: And out of that process, um the Gates Foundation, together with the government of UAE, created something called Aim for Scale.
- 8:37SPEAKER_01: Um, and that was a a new institution, um, and they each uh contributed to that, and that was designed to try to make progress, pushing these these ideas forward.
- 8:49SPEAKER_01: And then within those ideas, they picked one to start with, and that was one that also the Innovation Commission had been very excited about, which was using AI techniques for weather forecasting and trying to ensure that the benefits of that reached farmers, even in low and middle income countries.
- 9:09SPEAKER_01: So just to say a little bit more about that, this is a major dilemma for farmers is um is if you're thinking in you and I are both used to uh uh the seasons that we get in the temperate climates.
- 9:24SPEAKER_01: Um if you're in tropical climates, a key distinction is rainy season and dry season, um, you know, rather than summer, winter, spring, fall.
- 9:33SPEAKER_01: Um and the um if you're a farmer, if you're a farmer in rain-fed agriculture and you don't have irrigation, a key issue is when do you plant?
- 9:43SPEAKER_01: And as a rough rule of thumb, you typically want to start planting as soon as the rainy season has arrived, so you get the maximum growing season.
- 9:51SPEAKER_01: But if you plant too soon because you've seen some rain, but it's not really the rainy season, then that's a disaster because your seeds won't germinate.
- 10:00SPEAKER_01: If you've applied some inputs with the seeds, you may lose them.
- 10:04SPEAKER_01: That creates a really tricky situation for farmers.
- 10:07SPEAKER_01: Well, the good news is that weather forecasting is really advancing.
- 10:12SPEAKER_01: And that's one example.
- 10:13SPEAKER_01: There are many examples.
- 10:15SPEAKER_01: I was just read a story a day ago about uh chili farmers in India.
- 10:20SPEAKER_01: When they harvest their chilies, they have to leave them out to dry.
- 10:23SPEAKER_01: But if it rains, their crop's gonna rot.
- 10:26SPEAKER_01: The women who are farming these chilies have to, they can't really sleep at night because they have got to keep one ear open.
- 10:31SPEAKER_01: They've got to, it's you know 95 degrees, they've got to turn off their fan because if there's rain, they have to run out there and throw a tarp over the chilies, right?
- 10:40SPEAKER_01: Lots of circumstances where even a basic weather forecast would help you a lot.
- 10:45SPEAKER_01: Um, I was starting to say the good news.
- 10:47SPEAKER_01: The good news is AI weather forecast state over the past few years tremendous progress.
- 10:53SPEAKER_01: It went from not really being in contention and people thought, okay, it's a decade away, to now, for most applications, not all of them, being probably more accurate and longer lead times than traditional weather forecasting.
- 11:08SPEAKER_02: Let me stop you here for just a second.
- 11:10SPEAKER_02: When I think of weather forecasting, I think of a of an individual in front of a fake screen telling me that it's going to be windy and 65 degrees tomorrow.
- 11:20SPEAKER_02: But you're talking about AI.
- 11:21SPEAKER_02: And when I think about AI, I think large data sets and information being churned and something spits out and tells you something.
- 11:27SPEAKER_02: Is it when you talk about AI weather forecasting, is that what you're talking about?
- 11:30SPEAKER_02: I mean, how does it work?
- 11:32SPEAKER_02: Yeah.
- 11:33SPEAKER_01: Great, great.
- 11:34SPEAKER_01: So there are two elements, and I'm right now happy to talk about the other one as well.
- 11:39SPEAKER_01: There's the AI for the actual science, and then there's the AI for communications.
- 11:43SPEAKER_01: So for now, I'm talking about the AI for the actual science.
- 11:46SPEAKER_01: So, and I'm an economist, I'm not a weather forecaster.
- 11:49SPEAKER_01: There's a weather forecaster two doors down from me, so I can uh can bring him in if you want a better answer.
- 11:55SPEAKER_01: But basically, the way that traditional weather forecasting models work, you divide up the world into small grid cells.
- 12:03SPEAKER_01: In each grid cell, you know the existing weather conditions, and then you use the laws of physics to say how is each cell going to influence the next cell if there's a cold fronts coming in, things like that.
- 12:15SPEAKER_01: And then you run that model on supercomputers and you project forward the weather like that.
- 12:20SPEAKER_01: Now, and that's obviously brought tremendous progress to the world, the invent the development of those techniques, and they're still very valuable.
- 12:28SPEAKER_01: Um, but a what what supercomputers do is they look at, sorry, what the that traditional project approach requires supercomputers.
- 12:37SPEAKER_01: What AI approaches do is they look for patterns in the data of what's going to predict whatever phenomena you're you're interested in, and they use that type of information as well.
- 12:47SPEAKER_01: And I don't want to draw too strong a contrast if you either do a physics-based approach or you do a AI-based approach.
- 12:55SPEAKER_01: You know, some of the most exciting approaches are really combining elements of physics and elements of AI.
- 13:01SPEAKER_01: But one of the features is once you've got an AI model, it can then be run on a on a desktop.
- 13:07SPEAKER_01: It's it's they run 80,000 times faster than the traditional physics-based uh models with all of these grid cells projected out going forward.
- 13:16SPEAKER_01: Those uh that means that they can be tailored for particular forecasts that are needed by farmers.
- 13:22SPEAKER_01: Because you know, if you're if you're a typical office worker in the US, maybe you need to know do I bring a, you know, do I bring an umbrella or what's the 15 minutes I can ride my bike and avoid the rain or something.
- 13:34SPEAKER_01: If you're a farmer, you might want to know not just when did the rain start, but when are the when are continuous rains starting, because if you have a dry spell, that could kill off your crops as well, and you wouldn't want to plant.
- 13:46SPEAKER_01: So you can feed in whatever it is in particular that will be important to farmers, or for that matter, other users, and you can develop a uh a model that's targeted towards that, and that's can be done much more inexpensively than with traditional models, which means yeah.
- 14:04SPEAKER_02: Does the farmer have to put the the circumstances that they're interested in and knowing about, or is that something that somebody else says and the farmer can click on an app and get get that information?
- 14:15SPEAKER_02: How does the farmer get that information?
- 14:17SPEAKER_01: Well, let me give you a concrete example with in in the case of India.
- 14:21SPEAKER_01: So and I'll say what India is doing now, but I'll also say what India, you know, what what the aspirations would be for the future.
- 14:28SPEAKER_01: In 2025 year, the month the rainy season in India is the monsoon.
- 14:33SPEAKER_01: And um the um and that's vitally important for Indian farmers and and for many other people in India as well.
- 14:42SPEAKER_01: And the Indian government distributed forecasts to 38 million farmers that were and they sent those out with a for now, they sent them out with a fairly traditional technology, just uh a text message, SMS message to farmers.
- 14:57SPEAKER_01: And they also used WhatsApp a bit as well.
- 14:59SPEAKER_01: They also used in some areas these uh what's called interactive voice recording.
- 15:04SPEAKER_01: So there's a uh audio message that's played.
- 15:07SPEAKER_01: But obviously India's very, you know, making a lot of progress on AI and as an AI leader, and they're now launching um a whole farmer advisory system, which will go well beyond just weather, um, but will be a large language model that farmers interact with, the way you'd interact with uh Chat GPT or or Claude, um, and you can have a more natural conversation with it, and the farmer can say things like, I'm growing this particular crop and I'm in this particular district, and what advice would you have for me?
- 15:41SPEAKER_01: That's the long-run aspiration.
- 15:43SPEAKER_01: But what they did in 2025 was they just used the science to forecast something that was that the Indian government knew was important to a lot of farmers, the arrival of the monsoon, and they uh and they they distributed that with traditional text messages.
- 15:59SPEAKER_01: So that's why I'm separating out the forecast generation where AI can be applied, and then the communication.
- 16:05SPEAKER_01: Yes, you could use large language models for that.
- 16:08SPEAKER_01: I think that that's really a second area of digital agriculture and use of AI in that, which has many possible applications, whether being one of them advice on pests, soil, chemistry information.
- 16:20SPEAKER_01: A farmer in the US can afford to get their own, they might have thousands of acres.
- 16:25SPEAKER_01: They can certainly get soil testing done on their on their farm.
- 16:28SPEAKER_01: But if you're a very have a if you have a one-acre farm and you're living in a very low-income country, you might not be able to do that.
- 16:37SPEAKER_01: But the government might be able to have general soil testing done and give you the the soil composition, chemical soil composition for your area.
- 16:48SPEAKER_01: And that's if we think about the current the time we're in right now, fertilizer prices are spiking because of the war in in the Middle East.
- 16:58SPEAKER_01: And so it's really important that farmers understand what fertilizer they need and what fertilizer they don't need.
- 17:04SPEAKER_01: And if you have the good soil chemistry information, then you can start to optimize your fertilizer mix.
- 17:10SPEAKER_01: So that would be another um another use of digital agriculture and you know, large language models and AI that are set up to communicate with farmers, that's in a totally different area than weather.
- 17:23SPEAKER_02: Trevor Burrus, Jr.: But based on your experience both in the lab and based on your experience with the commission and your affiliation with AIM for scale, what two or three technologies do you think hold out the most promise or innovations hold out the most promise for significantly either improving the either the productivity or the sustainability of what's being produced in terms of food?
- 17:45SPEAKER_01: So let me break this out a little bit by the time scale.
- 17:48SPEAKER_01: I think one of the really striking things was that there's some low-hanging fruit that could be delivered very, very soon, or could be delivered already.
- 17:57SPEAKER_01: And when I talked about using AI weather forecasts, India is already doing that.
- 18:01SPEAKER_01: I think other, you know, Aim for Scales and uh and uh and the University of Chicago and and an organization called Precision Development.
- 18:09SPEAKER_01: We're working with the government of Ethiopia on that.
- 18:12SPEAKER_01: I think m I think many, many countries around the world could do that soon.
- 18:16SPEAKER_01: And I think that would we estimate the benefits.
- 18:19SPEAKER_01: Well, I say we estimate them.
- 18:21SPEAKER_01: There's trials around the world, farmers react to weather forecasts if you give them the information.
- 18:26SPEAKER_01: They use that information.
- 18:28SPEAKER_01: And we estimate the gains at, you know, very rough estimate, easily, you know, can easily be like 100 to 1 benefit cost ratio.
- 18:36SPEAKER_01: So that's uh that's something that could be done right away.
- 18:39SPEAKER_01: Longer run, I think there's you know, really tremendous potential.
- 18:43SPEAKER_01: I I talked about digital agriculture more broadly and uh and use of AI for I think that's another one that um in some areas like weather could be done right away.
- 18:53SPEAKER_01: Other areas might take a little bit longer, but it's is not far away at all.
- 18:57SPEAKER_01: I do think there's tremendous opportunities from advances in biology as well.
- 19:03SPEAKER_01: We've seen really fundamental advances in biological sciences, and those have been so far, there's a lot of work applying those to health, obviously, and there's some some very exciting work applying that to agriculture, but I think much more could be done.
- 19:19SPEAKER_01: I think you if we if we're thinking about environmental and climate challenges, I'll give one example.
- 19:25SPEAKER_01: I'm not saying I I don't necessarily want to say this is the example, but I think it's illustrative.
- 19:31SPEAKER_01: It may be uh a decent share of of emissions that cause climate change is methane, including methane from cows.
- 19:40SPEAKER_01: It may be possible to develop a vaccine that would really dramatically cut methane emissions.
- 19:46SPEAKER_01: I think that illustrates, I think that's an interesting case because that's something that could be socially very beneficial, but it's not something that there's a huge commercial incentive to invest in right now.
- 19:59SPEAKER_01: Yeah.
- 19:59SPEAKER_01: There might are farmers going to pay for that vaccine unless they can get a higher price for their milk by using uh by by you know having their cows vaccinated for it.
- 20:09SPEAKER_01: And yeah, there's probably some niche markets if some consumers would pay more for milk that uh that was you know had reduced methane emissions with it.
- 20:20SPEAKER_01: But if you really, you know, that's a limited the commercial incentives for that are very small relative to the potential social benefits.
- 20:28SPEAKER_01: And that's an area where I think there could be a role for philanthropy, for governments, for international institutions in trying to create incentives for development of something like a methane vaccine.
- 20:39SPEAKER_02: Last year, the Globe Global Dairy Platform Group, uh, which represents the major processors and major co-ops, dairy co-ops around the world, convened a number of interested parties from the food industry, from agriculture to discuss how they could tap into quote unquote the carbon markets, the financial mechanisms to create the revenue necessary to encourage farmers to embrace these practices.
- 21:06SPEAKER_02: Significant barriers currently exist because of the inability to measure and quantify and verify the results of particular strategies, whether it's land management or whether it's enzymes that reduce methane, et cetera.
- 21:20SPEAKER_02: So I think there is work going underway, and they they did receive a fairly significant grant from a development organization and development bank in Africa to basically work on uh a major project uh among several countries' dairy farmers.
- 21:34SPEAKER_02: You know, one of the challenges is getting the smallholder farmer to even know about any of this.
- 21:39SPEAKER_02: I can remember when I was governor of Iowa traveling to South Africa and Nigeria with the then Ambassador Andrew Andy Young.
- 21:47SPEAKER_02: Um, and the ambassador and I were basically trying to talk to farmers about the necessity of increasing productivity, the use of biotechnology, and things of that nature.
- 21:57SPEAKER_02: And I will never forget talking to one uh African Farmer, I think he was in South Africa.
- 22:01SPEAKER_02: And it was like we t we we told him what what he could do.
- 22:04SPEAKER_02: He could plant this and it would increase his productivity.
- 22:07SPEAKER_02: And but he said, But my grandfather and my great grandfather, they've always done it this way, and that's the way I'm gonna do it.
- 22:14SPEAKER_02: How do you how do you change the mindset?
- 22:17SPEAKER_02: I mean, you've studied change, you've you you you've been engaged, you've be you're a Nobel Prize winner uh as a result of some of the things that you've done on change.
- 22:25SPEAKER_02: How do you how do you do that?
- 22:26SPEAKER_02: How do you get people to change?
- 22:28SPEAKER_02: Maybe that's a a fairly big question, but but how how do you do it?
- 22:33SPEAKER_02: What needs to be done?
- 22:34SPEAKER_01: Right.
- 22:35SPEAKER_01: Well, I think a key thing is to focus on the needs of the farmers.
- 22:40SPEAKER_01: And to start, you know, I don't want to claim it's impossible to change, you know, uh farmer attitudes, but I think you need to start with the needs of the farmers and what the farmers want.
- 22:50SPEAKER_01: And so, you know, I think one of the reasons why we went through that process of what the commission decided to focus on, that's why we wanted evidence of impact and cost effectiveness, say for the new seeds, but we also wanted evidence that farmers would respond to that.
- 23:07SPEAKER_01: And that's one of why one of one of the reasons why one of our top innovations was these improved weather forecasts, because in randomized trials where there's a treatment group and a comparison group, where the treatment group got new weather forecasts and the uh and the comparison group uh didn't have access to those.
- 23:26SPEAKER_01: We saw changes in farmer behavior.
- 23:28SPEAKER_01: I say we, I'm summarizing the literature here, so it's a half dozen studies around the world, but in virtually all of them they found farmers responded.
- 23:36SPEAKER_01: So that's I'd rather start with something like that than to take the opposite example, the methane vaccine.
- 23:43SPEAKER_01: That's something where it's going to be hard to get farmers to adopt unless there's a financial incentive for them.
- 23:49SPEAKER_01: And it's not going to be easy to create, I would actually think that a methane vaccine probably will take off first among farmers with lots of lots of livestock in higher income countries where you it's easier to set up that monitoring and verification system and so that you can't tap carbon uh finance, for example.
- 24:09SPEAKER_01: If you can do it in in in uh in Nigeria as well, then great.
- 24:14SPEAKER_01: Maybe we should be thinking about maybe the first place this will happen will be California or Wisconsin or the Netherlands, and and then it'll diffuse from there.
- 24:24SPEAKER_01: But let me answer your question because you've uh I've been ducking it so far and saying, well, just plan work around the needs of the farmer.
- 24:31SPEAKER_01: Let me take the digital agriculture example.
- 24:33SPEAKER_01: So this is an area where I've been working for a while, and one thing we found is that in country after country, the agriculture experts in the Ministry of Agriculture, they'll come up with good, technically sound agricultural advice.
- 24:50SPEAKER_01: And but if you just try to take the raw advice that they provide and send it out through a text message or something like that to farmers, it's not necessarily put in a way that farmers can understand and respond to.
- 25:03SPEAKER_01: And there's some cases where you get very technical messages about depending on what the pH level of your soil is, you should do this or do that.
- 25:12SPEAKER_01: Well, yeah, a typical farmer in some, you know, in some low-income areas will not know what pH is and certainly have no way of affording a test of their soil to to test for pH or know how to how to conduct it.
- 25:27SPEAKER_01: So I think this is a case where what is important is to do what what a tech company would do, which would be to try out different types of messages.
- 25:37SPEAKER_01: You know, it's a sort of a mixture of you need people who are very close to the farmer who are used to dealing with farmers.
- 25:43SPEAKER_01: You need people who can set up an A-B test on try different messages.
- 25:48SPEAKER_01: You know, then and you should talk to extension workers, to farmers organizations, to farmers to try and craft messages.
- 25:56SPEAKER_01: But then you through a process of talking to farmers, trying out different messages, eventually you come up with better and better messages, and then farmers are more likely to respond to them.
- 26:06SPEAKER_01: And I found, I mean, I think there can be a lot of cynicism about governments, and in some ways, the story I've just told, you could read that as saying, you know, the glass is half empty, the governments are getting it all wrong, they're not developing good messages.
- 26:20SPEAKER_01: But I think the good news is in country after country, whether it's Kenya or Ethiopia, um, or Act Ministries are pretty open to that evidence.
- 26:30SPEAKER_01: They don't necessarily have, they don't um they don't necessarily have the people on staff who who are going to be able to do that.
- 26:38SPEAKER_01: But um, you know, once they see the evidence, they're willing to do things like I'll give you an example from Ethiopia.
- 26:45SPEAKER_01: But there they had a voice act a voice system.
- 26:47SPEAKER_01: So press one for this information, press two for that information.
- 26:51SPEAKER_01: It started off with press one if you want information from before, you know, for as you plan your planting and all the way through the agricultural season.
- 27:00SPEAKER_01: Well, a very simple thing is change the order so that the farmer, instead of having to list list one through eight before you get press nine for this, if it's a season where nine is relevant, start with number nine.
- 27:12SPEAKER_01: Um, you know, those are you know basic tricks and tips like that can go a long way.
- 27:18SPEAKER_01: And just you know, the the overall approach, which uh we try to follow in the Development Innovation Lab and I think other development economists have have uh found is you don't you can have your theories about what will be the best way to communicate with farmers, but you really have to test them out.
- 27:35SPEAKER_01: And that's that's that uh you know, there's a role for the theory, but until you actually have the evidence from some sort of experimental approach, you know, typically from some sort of experimental approach, you don't don't really know.
- 27:47SPEAKER_02: You know, I was a lawyer before I got involved in politics and and uh public service.
- 27:52SPEAKER_02: And I suppose there was some way in which my practice informed my personal life in terms of how I how I went about making decisions.
- 27:59SPEAKER_02: But I as I was listening to you, uh I just have to ask you this question.
- 28:03SPEAKER_02: As I understand the work that you've done, this is very I'm gonna make it incredibly simplistic, and I apologize for that, but it's the only way I can explain it.
- 28:10SPEAKER_02: It's as if, you know, it's the old adage, don't bite off more than you can chew.
- 28:14SPEAKER_02: You got a big problem.
- 28:16SPEAKER_02: Don't don't try to swallow the whole thing at once.
- 28:18SPEAKER_02: Take bits and pieces of it and try to figure it out and work your way through it.
- 28:22SPEAKER_02: Is the work that you've done for nations, for companies, for your partners, has it changed the way in which you make personal decisions?
- 28:29SPEAKER_02: Has it informed the way in which you go about making what would be for you a big personal decision differently?
- 28:36SPEAKER_01: Yeah, that's um that's that's that's a great question.
- 28:39SPEAKER_01: Well, on the first part of the question, yeah, I think a lot of what, you know, a lot of the approach that there are huge problems in international development.
- 28:48SPEAKER_01: I think we've seen I tend to be more on the optimistic side.
- 28:52SPEAKER_01: You know, we've seen tremendous reductions in poverty and hunger.
- 28:56SPEAKER_01: Yeah, I I've been working in I'm old enough, I've been working in this field for a while, and um I work in Kenya, I can see people are living much better lives than than they were when I first started working there.
- 29:09SPEAKER_01: And they've got a better standard of living, a lot more freedom because it was a one-party state when I when I first started working there.
- 29:16SPEAKER_01: They've got more sources of information, so much more educated.
- 29:20SPEAKER_01: Um, so I think there is a lot of progress, but I think a lot of that does come from you know tackling individual problems one, you know, each piece by piece.
- 29:30SPEAKER_01: Um I I don't I don't think we'll necessarily get one solution uh that fixes everything.
- 29:35SPEAKER_01: Um and then in terms of how do I how I live my own life, uh I mean, here's, you know, what look when we're when, you know, from a family perspective, if we're thinking about a medical decision, you know, we will try, you know, I'm very familiar with the basics of randomized trials in economics, and some of the statistics is gonna be somewhat similar to the w to the statistics used in health.
- 30:03SPEAKER_01: So um, you know, a very basic principle is I'm gonna trust evidence from a randomized trial more than I'll trust a raw correlation that, you know, much as I like, I've got a coffee cup here, you know, much as I drink coffee, I'm not primarily doing that because I've seen some study that coffee drinkers are more, you know, live longer or something like that.
- 30:26SPEAKER_01: If there was a randomized trial on that, I have no idea whether there is, you know, I'd be more likely to believe that evidence.
- 30:32SPEAKER_01: At that level, that's probably the most important, you know, that's obviously among the most important decisions that anybody makes in their life is is um about medical care.
- 30:41SPEAKER_01: And certainly our family will pay a lot of attention to evidence from randomized trials on that.
- 30:46SPEAKER_02: Aaron Powell Well, let me uh finish with one last question.
- 30:48SPEAKER_02: As you look at your crystal ball five years, ten years down the road, do you believe that we will have made progress in terms of food security and for all?
- 30:57SPEAKER_02: Do you think we'll be in relatively the same position, or do you think because of the changing climate and some of the challenges that we face with that, we might actually be faced with a an even larger challenge?
- 31:08SPEAKER_02: Are you hopeful or are you not?
- 31:10SPEAKER_01: Yeah.
- 31:11SPEAKER_01: Very tough question.
- 31:12SPEAKER_01: And unfortunately, um, you know, I'm I'm not not really in the crystal ball uh version of economics.
- 31:18SPEAKER_01: Uh uh, but I will, you know, nonetheless, let me try and tell you my my personal guess.
- 31:23SPEAKER_01: You know, we've got big chat challenges and uh the long run trend from when I began my career till now is huge improvements, reductions in in hunger.
- 31:33SPEAKER_01: Recent trend has not been.
- 31:34SPEAKER_01: It's been been negative.
- 31:36SPEAKER_01: And obviously, even in just you know, very, very recently, we've seen this spike in fertilizer prices.
- 31:42SPEAKER_01: And that's you know, that could be very bad for for a lot of farmers and for fruits eventually if it's sustained uh for food security.
- 31:50SPEAKER_01: And so, but if I think five years out, look, I'm none of us know what what the impact of AI is going to be, but I think it has the potential to be very useful for not just for the fortunes of people in Silicon Valley, but for ordinary people all around the world, and maybe even in particular for people in low and middle income countries.
- 32:15SPEAKER_01: If you could have a situation where farmers not only got good information on the weather, but they got good information on you know, if a pest comes, well, what type of pest is that?
- 32:26SPEAKER_01: What action should I take against that pest?
- 32:29SPEAKER_01: If there's a uh, we talked about soil chemistry and what what fertilizers or other inputs do I do I need if I if there's a new crop that looks like it might be profitable, but I'm not used to growing that before.
- 32:42SPEAKER_01: Well, I could watch videos that are are um are giving me advice on exactly how you care for this plant.
- 32:48SPEAKER_01: And maybe the videos are, you know, I could be recorded, I can show you what I'm doing and the it can give me coaching.
- 32:54SPEAKER_01: The same way, you know, I've tried I've tried to just try to, you know, start using, you know, doing some exercises at home.
- 33:02SPEAKER_01: And, you know, if the AI, you know, you know, video me and give me uh give me tips.
- 33:08SPEAKER_01: So I think there's huge, huge potential in that area.
- 33:12SPEAKER_01: Now, the problem is the companies are gonna naturally go after, and this is what I was trying to get after with the methane example, there's strong commercial incentives to develop some products using AI.
- 33:24SPEAKER_01: And that's great.
- 33:25SPEAKER_01: Those products should be developed, but there are other products that are also really important where the commercial incentives might be a little bit weaker.
- 33:32SPEAKER_01: And there, I think it's important for philanthropy, for governments to figure out how are we going to create either how are we going to invest in that RD or how are we going to create incentives for the private sector to invest in that RD.
- 33:45SPEAKER_01: I'll give one more example also from India.
- 33:49SPEAKER_01: India's now got software to help its community health workers do their jobs, and that's proven, you know, rigorously proven to lead to better, better outcomes for patients.
- 34:01SPEAKER_01: Well, that same thing should be there for agricultural extension agents all around the world.
- 34:06SPEAKER_01: But governments have to procure that software, and you're the one with the government experience.
- 34:11SPEAKER_01: But it's not easy to write a procurement contract for, you know, for software.
- 34:16SPEAKER_01: And so figuring out how to do that, that's where we need, I'll go back to your legal background.
- 34:22SPEAKER_01: You know, that's where you need people who know AI and computer science, people who know law, people, I think I'll I'll speak up for my profession, people know economics, all thinking about how should governments go about procuring AI solutions for a range of public services, but definitely including services for farmers.
- 34:41SPEAKER_02: Well, this has been a fascinating conversation.
- 34:44SPEAKER_02: Really appreciate your time.
- 34:46SPEAKER_02: Uh, I wish we had more time, and we might bring you back for a second round.
- 34:50SPEAKER_02: Really appreciate the work that you're doing, and certainly look forward to learning more about the innovations uh that Aim for Scale is focused on.
- 34:58SPEAKER_02: So, Professor, thank you very much for uh taking a few minutes.
- 35:01SPEAKER_00: Thank you.
- 35:03SPEAKER_00: Thank you for listening to the World Food Prize Podcast.
- 35:06SPEAKER_00: If today's conversation inspired you, we invite you to take the next step.
- 35:11SPEAKER_00: By making the gift, you help support dialogue like this, elevate innovative solutions, give voice to science, and invest in the next generation of leaders working to end hunger.
- 35:24SPEAKER_00: Visit WorldFood Prize.org to donate, learn more, and get involved.
- 35:30SPEAKER_00: Together, through dialogue and action, we can sustainably feed the world.