WEBVTT - NASA and AI: Decoding Our Universe 0:00:02.080 --> 0:00:04.960 If I were to go back, I don't know thirty 0:00:05.040 --> 0:00:09.200 years in Kenya, what's the difference between then and now 0:00:09.280 --> 0:00:12.840 in terms of tree cover. I'm talking to Philip, Theigo 0:00:13.440 --> 0:00:16.160 Special Technology envoid to the Kenyan President. 0:00:16.760 --> 0:00:19.400 Let's speak as if you think about we are now 0:00:19.400 --> 0:00:22.759 elevent trol percent. Previously we were more than twenty percent. 0:00:23.160 --> 0:00:26.560 So we are cutting trees more than we're planting them. 0:00:26.840 --> 0:00:30.560 In thirty years, Kenya lost half its tree cover half 0:00:31.280 --> 0:00:34.760 And here's why that matters. Kenya is a mountainous country. 0:00:35.280 --> 0:00:38.120 Dotted throughout the highlands are dozens of what canyons call 0:00:38.280 --> 0:00:44.480 water towers natural reservoirs, densely forested areas capable of absorbing 0:00:44.520 --> 0:00:46.760 the enormous amount of water that falls on the country 0:00:46.840 --> 0:00:51.400 during the rainy seasons. The tree roots and undergrowth secure 0:00:51.520 --> 0:00:55.200 and capture moisture, then slowly release it into the rivers 0:00:55.200 --> 0:00:58.560 that flow down into the country's low lying coastal areas. 0:00:59.160 --> 0:01:04.080 But in recent years, the water towers have depleted, settlements 0:01:04.240 --> 0:01:08.080 have encroached on them, trees have been chopped down, thousands 0:01:08.080 --> 0:01:11.800 of acres cleared. The natural reservoirs cease to hold nearly 0:01:11.840 --> 0:01:15.360 as much water. So now Kenya is prone to extremes. 0:01:15.800 --> 0:01:18.319 Too much water flowing down from the highlands in the 0:01:18.400 --> 0:01:21.880 rainy season and too little water left during the dry season. 0:01:22.600 --> 0:01:24.640 So you have a couple of hours of water, then 0:01:24.680 --> 0:01:27.000 you have a couple of hours with no water that 0:01:27.080 --> 0:01:29.840 it tops off to be dry by the city authority. 0:01:30.040 --> 0:01:33.120 So that's the significance of the water towers we have 0:01:33.200 --> 0:01:34.360 when they cannot hold water. 0:01:34.800 --> 0:01:39.000 Kenya desperately needed to restore its water towers by planting 0:01:39.319 --> 0:01:42.760 as many trees as humanly possible. So in the fall 0:01:42.800 --> 0:01:46.679 of twenty twenty three, the Kenyan government took action. It 0:01:46.760 --> 0:01:51.360 started a national holiday, National Tree Growing Day, a day 0:01:51.400 --> 0:01:54.120 to allow the citizens of Kenya to go out into 0:01:54.120 --> 0:01:57.480 the forest to dominate the Kenyan countryside and plant as 0:01:57.520 --> 0:02:02.200 many trees as they can. Government decided on a number. 0:02:03.080 --> 0:02:08.040 The President's really focus right around how to ensure that 0:02:08.080 --> 0:02:10.440 we do not lose more forests was in this very 0:02:10.440 --> 0:02:12.800 ambitious campaign around fifteen billion trees. 0:02:13.320 --> 0:02:16.840 That's right, fifteen billion with a bee. 0:02:17.360 --> 0:02:19.679 So imagine that number will tell you the ambition. But 0:02:19.760 --> 0:02:22.400 as he tells you the deficit, it has to be 0:02:22.440 --> 0:02:24.720 fifteen billion in the next eight years. 0:02:25.080 --> 0:02:28.320 Fifteen billion trees over eight years averages out to more 0:02:28.360 --> 0:02:32.440 than five million trees per day. That's a lot of trees. 0:02:33.000 --> 0:02:35.560 But with such a massive goal, how can you track 0:02:35.639 --> 0:02:38.320 your progress? How do you know where to plant those 0:02:38.320 --> 0:02:41.480 trees so they'll have the most impact. How do you 0:02:41.560 --> 0:02:45.480 monitor where older trees are still being cut down? Well, 0:02:45.520 --> 0:02:49.200 the answer to those questions came from IBM and a 0:02:49.240 --> 0:02:55.120 little space agency called NASA. That's right, folks, Smart Talks 0:02:55.800 --> 0:03:00.359 is going to space. My name is Malcolm Glabo listening 0:03:00.400 --> 0:03:04.040 to the latest episode of Smart Talks with IBM, where 0:03:04.040 --> 0:03:06.760 we offer our listeners a glimpse behind the curtain of 0:03:06.800 --> 0:03:10.800 the world of technology. In this season, IBM has gone 0:03:10.840 --> 0:03:15.840 inside elementary school classrooms, toured formulation labs at Loreel, and 0:03:15.919 --> 0:03:19.600 spoken with the fan development team at Scuderia Ferrari HP. 0:03:20.639 --> 0:03:25.160 In this episode, how IBM is partnering with NASA to 0:03:25.200 --> 0:03:29.720 build geospatial models using data from satellites to better understand 0:03:29.760 --> 0:03:35.640 our Earth and Solar system. 0:03:35.680 --> 0:03:43.840 Five four three two one zero all engine running liptoff. 0:03:44.160 --> 0:03:47.240 We have a liptof thirty two minutes past the hour 0:03:47.600 --> 0:03:48.960 liftoff on Apollo eleven. 0:03:49.680 --> 0:03:53.320 IBM has worked on space related projects since before I 0:03:53.400 --> 0:03:54.000 was even born. 0:03:54.680 --> 0:03:56.440 I'm all for man. 0:03:57.480 --> 0:04:00.760 A team of four thousand IBM engineers help create the 0:04:00.800 --> 0:04:03.600 Saturn five rocket that took Neil Armstrong to the Moon 0:04:05.600 --> 0:04:10.120 up and when I think of NASA, I tend to 0:04:10.160 --> 0:04:12.760 picture the Moon landing, or the team of people back 0:04:12.760 --> 0:04:16.640 in Houston guiding the Apollo mission, or the Hubble telescope, 0:04:16.920 --> 0:04:21.120 or astronauts aboard the International Space Station. What I didn't 0:04:21.120 --> 0:04:24.880 think about until now are NASA's geographers. 0:04:25.680 --> 0:04:27.599 In order to go places, you need a map things. 0:04:28.040 --> 0:04:32.120 This is Kevin Murphy, chief Science Data Officer at NASA's 0:04:32.240 --> 0:04:33.560 Science Mission Directorate. 0:04:34.320 --> 0:04:37.320 But I think that there's an assumption that NASAs all 0:04:37.360 --> 0:04:41.159 about rockets and astronauts, and certainly that's a really large 0:04:41.160 --> 0:04:42.480 part and important part of NASA. 0:04:43.240 --> 0:04:46.640 NASA sends people to space and looks out of the stars, 0:04:47.120 --> 0:04:50.680 but NASA also looks down at the Earth. The agency 0:04:50.720 --> 0:04:57.599 has about one hundred and fifty satellites that use radar, lightar, landset, Aquaterra, cloudset, AURA, 0:04:58.000 --> 0:05:02.960 low Earth orbit, Medium Earth orbit, geostationary orbit, on and on. 0:05:03.600 --> 0:05:07.760 In one sense, NASA makes hardware to build rockets and 0:05:07.800 --> 0:05:11.280 spacecraft and all those satellites that circle the Earth. But 0:05:11.400 --> 0:05:17.599 fundamentally NASA also collects data. It's scientists and engineers people 0:05:17.640 --> 0:05:21.160 like Kevin want to make the best use possible of 0:05:21.200 --> 0:05:25.240 all the information gathered by all those many dozens of instruments. 0:05:25.920 --> 0:05:29.599 Right now, we gather around twenty five petabytes of new 0:05:29.760 --> 0:05:32.680 observational data per year. In the next couple of months, 0:05:32.760 --> 0:05:38.279 we're about to launch a high resolution global radar. When 0:05:38.279 --> 0:05:41.840 that launches, will double how much we collect every year 0:05:41.960 --> 0:05:44.000 to about fifty petabytes of information. 0:05:44.680 --> 0:05:49.080 Actually, since we recorded this conversation, NASA launched that global radar, 0:05:49.279 --> 0:05:53.640 what they call NYSAR. So NASA is already generating new 0:05:53.680 --> 0:05:57.160 data at the rate of fifty petabytes each year. To 0:05:57.160 --> 0:06:00.679 put that in perspective, a single petabyte could hold about 0:06:00.720 --> 0:06:05.280 five hundred billion pages of standard printed text. Now can 0:06:05.480 --> 0:06:07.960 anyone sort of apply to use this data is. 0:06:08.279 --> 0:06:11.039 They don't even have to apply. It's free and open data. 0:06:11.160 --> 0:06:14.960 It advances how we understand what we do on Earth 0:06:15.000 --> 0:06:18.840 and how we see ourselves within the universe. People can 0:06:18.880 --> 0:06:21.680 take it for so many different downstream applications. So you 0:06:21.720 --> 0:06:24.560 can go to our websites today, you can search through 0:06:24.600 --> 0:06:29.000 our tools, and you can download information from the Mars rovers. 0:06:29.040 --> 0:06:32.720 You can download information from the lunar reconnaissance orbiter or 0:06:32.760 --> 0:06:34.920 any of the science data satellites, and. 0:06:34.880 --> 0:06:38.919 Give me an example of a really cool application, a 0:06:38.960 --> 0:06:41.640 really cool use that someone I don't know in academic 0:06:41.760 --> 0:06:44.080 whatever has used your data for. It is there? 0:06:44.120 --> 0:06:47.360 It okay. So one of the really kind of cool 0:06:47.440 --> 0:06:51.320 but unexpected observations that we had is that we launched 0:06:51.360 --> 0:06:55.720 a pair of satellites in their early two thousands called Grace, 0:06:56.120 --> 0:06:59.000 and these satellites orbit the Earth and they can measure 0:06:59.080 --> 0:07:01.800 very precisely the distance that they're away from each other 0:07:01.839 --> 0:07:04.520 as they orbit the Earth, and as you go into 0:07:04.560 --> 0:07:07.960 gravity wells, you can actually see a satellite accelerate and 0:07:08.000 --> 0:07:11.720 the other one accelerate after it. Right, And using that information, 0:07:12.160 --> 0:07:14.840 we were trying to map kind of the gravity fields 0:07:15.120 --> 0:07:17.880 of Earth. What they found is that they can actually 0:07:17.960 --> 0:07:21.600 map below kind of the mass of Earth to where 0:07:21.680 --> 0:07:25.240 water storage is. For instance, so aquifers, right, so you 0:07:25.240 --> 0:07:30.400 can monitor through gravity how much water is being depleted 0:07:30.520 --> 0:07:33.840 or added to an aquifer or the density of glaciers. 0:07:34.440 --> 0:07:37.760 So, just to back up for a moment, the presence 0:07:37.960 --> 0:07:42.280 and density of water deposits below the Earth's surface have 0:07:42.360 --> 0:07:47.280 an effect on gravitational fields that are being measured in 0:07:47.320 --> 0:07:51.480 space correct. Yeah, And so does that tell you presume 0:07:51.720 --> 0:07:53.760 you learn things like where there's an aquifer where you 0:07:53.760 --> 0:07:55.600 didn't think there was an aquifer. 0:07:55.640 --> 0:07:58.840 Or if it's being depleted faster. Yeah. Yeah. 0:07:59.240 --> 0:08:01.320 So who's using that kind of data? 0:08:01.800 --> 0:08:05.320 All sorts of different organizations, whether they're you know, NGOs 0:08:05.600 --> 0:08:08.920 or government agencies or people that are planning a large 0:08:08.960 --> 0:08:09.960 agricultural product. 0:08:10.040 --> 0:08:12.160 How did you Was that an intentional decision? 0:08:12.200 --> 0:08:13.520 It wasn't. It was accidental. 0:08:14.200 --> 0:08:20.800 It was accidental. NASA has assembled a historically unprecedented mountain 0:08:20.800 --> 0:08:24.640 of data about the physical world, free and open to anyone, 0:08:25.040 --> 0:08:27.800 and the possibilities for how that information can be used 0:08:27.840 --> 0:08:33.840 are so vast that even NASA is still uncovering them. 0:08:34.040 --> 0:08:36.760 When I was a kid, I loved legos. I had 0:08:36.760 --> 0:08:39.880 a huge bin full of them. At the time, legos 0:08:39.960 --> 0:08:43.480 were really just colored bricks of various sizes. They weren't 0:08:43.520 --> 0:08:46.440 as complicated as they are today. And what I realized 0:08:46.480 --> 0:08:49.040 even then was that there were more possibilities in a 0:08:49.040 --> 0:08:51.960 box of legos than I could ever imagine on my own. 0:08:52.760 --> 0:08:54.440 I played with my brother and he would show me 0:08:54.480 --> 0:08:56.720 something that hadn't occurred to me, and I go to 0:08:56.760 --> 0:08:59.080 my friend Bruce's and see that he was off on 0:08:59.120 --> 0:09:01.920 some legos. Tan engined that I'd never even thought of, 0:09:02.480 --> 0:09:04.520 like a cool bridge, or a castle or a truck. 0:09:05.280 --> 0:09:08.960 I use legos one way. Bruce used his legos in 0:09:09.000 --> 0:09:14.520 a completely different way. NASA's data treasure trove is like 0:09:14.559 --> 0:09:18.040 a very very big box of legos. And here's the question. 0:09:18.640 --> 0:09:23.760 With so much data, containing so many possible connections, could IBM, 0:09:24.080 --> 0:09:29.960 and specifically IBM's artificial intelligence help NASA scientists uncover patterns 0:09:29.960 --> 0:09:33.360 and connect systems in a way they've never done before. 0:09:35.840 --> 0:09:38.280 Everything started with a question, right. 0:09:38.400 --> 0:09:42.520 I'm talking to one Bernabe Moreno, director of IBM Research 0:09:42.559 --> 0:09:43.120 in Europe. 0:09:43.840 --> 0:09:47.720 As we advance AI, we have new tools to understand 0:09:48.520 --> 0:09:52.120 the surroundings, understand the world, understand the language, and understand 0:09:52.120 --> 0:09:55.199 our planets. And the question that we were asking ourselves 0:09:55.400 --> 0:09:57.880 was all these new advances that we see in language. 0:09:57.960 --> 0:10:01.280 It was a post GPT moment. Could we apply the 0:10:01.320 --> 0:10:05.000 same idea on the same architecture and technology to add 0:10:05.000 --> 0:10:06.200 a double do our planets? 0:10:06.760 --> 0:10:10.120 The advent of AI created a new opportunity. What if 0:10:10.120 --> 0:10:13.680 all of NASA's mountain of data could be organized, analyzed, 0:10:14.000 --> 0:10:19.280 understood by artificial intelligence. The original idea was to create 0:10:19.360 --> 0:10:23.000 a geospatial foundation model for the Earth and from there 0:10:23.440 --> 0:10:28.160 create additional specialized models for other scientific priorities of NASA, 0:10:28.920 --> 0:10:32.640 and finally create an AI system that can understand all 0:10:32.720 --> 0:10:36.480 the data across those specialized models in order to uncover 0:10:36.559 --> 0:10:41.720 hidden insights and relationships. Together, these models could unlock an 0:10:41.920 --> 0:10:46.480 infinite number of potential applications. I asked Kevin Murphy at 0:10:46.559 --> 0:10:49.840 NASA about the beginning of these Earth models. 0:10:50.440 --> 0:10:53.240 Has some colleagues, and we were investigating a number of 0:10:53.320 --> 0:10:58.480 different avenues of using AI with our data, but also 0:10:58.760 --> 0:11:01.360 kind of the management and stewardship of the data, so 0:11:01.480 --> 0:11:03.600 not only like the observations, but how we make it 0:11:03.640 --> 0:11:07.320 available to people, make it discoverable. And they said, hey, 0:11:08.000 --> 0:11:10.520 we see these transform architectures. We think that they can 0:11:10.600 --> 0:11:14.760 be applicable to some of the sequential observations that we make. 0:11:15.280 --> 0:11:17.400 We'd really like to work with IBM on that. And 0:11:17.440 --> 0:11:21.360 I was like, I'm really skeptical. But while because I 0:11:21.400 --> 0:11:27.480 hadn't seen those types of tools really produce results that 0:11:27.520 --> 0:11:31.520 were commensurate with the amount of effort you put into them, right, 0:11:31.600 --> 0:11:33.920 So we were getting some really good results and deep 0:11:34.000 --> 0:11:37.840 learning approaches, but they took a lot of effort. 0:11:37.520 --> 0:11:39.120 But Kevin came around quickly. 0:11:40.000 --> 0:11:44.680 When we typically develop a new data product or an algorithm, 0:11:45.120 --> 0:11:48.800 it takes anywhere from you know, twelve months, eighteen months, 0:11:48.880 --> 0:11:54.280 twenty four months to go from data and hypothesis to 0:11:54.480 --> 0:11:58.440 results which is validated. We were able to get approximately 0:11:58.480 --> 0:12:03.560 the same precision for some well known types of benchmarks 0:12:03.920 --> 0:12:06.000 with and I think it was about four months, oh, 0:12:06.040 --> 0:12:07.080 instead of starting the work. 0:12:07.200 --> 0:12:11.360 Yeah, yeah, so it happened faster than you thought, much faster. 0:12:12.440 --> 0:12:16.120 In twenty twenty three, IBM and NASA launched a foundation 0:12:16.320 --> 0:12:21.120 model trained on NASA's harmonized landset sentinel to satellite data 0:12:21.480 --> 0:12:25.520 across the continental United States. They named the model Prithvi, 0:12:26.040 --> 0:12:30.120 the Sanskrit word for Earth. The first version of Prithvy 0:12:30.480 --> 0:12:34.480 used only Earth observation images and just that was enough 0:12:34.520 --> 0:12:38.560 to totally change Kevin's idea of what foundation models could do. 0:12:39.559 --> 0:12:43.480 But they didn't stop there. IBM and NASA were encouraged 0:12:43.640 --> 0:12:47.960 at how well Prithvy worked for Earth observation tasks, so 0:12:48.240 --> 0:12:51.079 they decided to create a more complex version of Prithvy 0:12:51.480 --> 0:12:55.559 that could understand whether and climate data. They hope this 0:12:55.640 --> 0:12:58.760 new version of PRITHV would allow researchers to answer new 0:12:58.840 --> 0:13:02.440 questions about the Earth, from short term weather forecasting to 0:13:02.559 --> 0:13:06.160 longer term climate effects. Imagine you have a map of 0:13:06.360 --> 0:13:11.319 all the different temperatures, pressures, clouds, rainfall, and more from 0:13:11.400 --> 0:13:15.640 around the globe. With this map, IBM and NASA could 0:13:15.720 --> 0:13:19.679 implement advanced tasks. They could track the formation of El 0:13:19.760 --> 0:13:22.400 Nino or predict how the path of a hurricane would 0:13:22.440 --> 0:13:26.640 change if the ocean temperature went up by half a degree. 0:13:26.960 --> 0:13:30.040 I will always remember this moment was when we created 0:13:30.080 --> 0:13:35.160 the Weather and Climate Foundational Molive. The senior methodologist of NASA, 0:13:35.640 --> 0:13:38.480 it was like, I cannot believe that it has changed 0:13:38.679 --> 0:13:40.800 the way I think about the AI and ever since 0:13:40.840 --> 0:13:43.120 he's been kind of preaching with this A sample. 0:13:43.960 --> 0:13:46.440 One and his team then took the model and decided 0:13:46.480 --> 0:13:50.559 to test it, really tested it. Took away ninety nine 0:13:50.600 --> 0:13:53.839 percent of the data points and ran the experiment again. 0:13:54.520 --> 0:13:56.560 What they were trying to figure out is if the 0:13:56.640 --> 0:13:59.680 model had learned enough about the basic principles of the Earth, 0:14:00.120 --> 0:14:03.040 the underlying physics of the way the planet works, to 0:14:03.160 --> 0:14:06.480 fill in the blanks on its own with just one 0:14:06.520 --> 0:14:10.120 percent of the original data, would it still be accurate 0:14:10.200 --> 0:14:16.760 in its predictions. What happened the model crushed it, so 0:14:17.000 --> 0:14:19.440 it was able to extrapolate on the basis of one 0:14:19.480 --> 0:14:23.160 percent of the data what the entire picture looked like. Yes, 0:14:24.600 --> 0:14:28.280 because pre learned everything right, Yeah, it learned the kind 0:14:28.280 --> 0:14:32.880 of principles of exactly. Yeah. Oh wow, that's very very impressive. 0:14:33.000 --> 0:14:35.560 So at that moment when when you realize you could 0:14:35.640 --> 0:14:40.360 do that, and just curious about your emotional I mean, 0:14:40.360 --> 0:14:41.800 did you jump up and down? What did you do? 0:14:42.200 --> 0:14:45.400 That's I was like, wow, it was a very emotional 0:14:45.440 --> 0:14:49.360 meeting because you know, having this person say now I'm 0:14:49.400 --> 0:14:52.920 convinced right, Yeah, it was kind of a quite a 0:14:52.920 --> 0:14:55.840 special moment. These moments make your life as a researcher. 0:14:57.280 --> 0:15:00.880 IBM and that's a launch prithe for Weather and Climate 0:15:01.040 --> 0:15:04.040 in twenty twenty four. And while IBM and as a 0:15:04.080 --> 0:15:08.040 scientist could use Prithvy to run interesting experiments, they were 0:15:08.080 --> 0:15:11.640 even more excited about how prithy could help people in 0:15:11.640 --> 0:15:19.920 the real world. So let's go back to Kenya Ambassador 0:15:20.040 --> 0:15:23.600 Philip Digo and the country's Great tree planting project. 0:15:24.480 --> 0:15:27.440 So on those initial months, there was a massive effort, 0:15:27.480 --> 0:15:30.240 including a couple of national holidays. 0:15:30.400 --> 0:15:31.440 For tree planting. 0:15:32.240 --> 0:15:35.760 Yes, where the entire cabinet was sent. 0:15:36.040 --> 0:15:38.640 Ah, did you plant trees as I did? 0:15:38.720 --> 0:15:40.600 Oh my god, I said, the entire cabinet plus someone 0:15:40.800 --> 0:15:41.600 we have to be seen. 0:15:42.040 --> 0:15:43.840 Are you good at that plantet two weeks ago? 0:15:44.200 --> 0:15:46.080 Well it's very easy to go hole put a tree 0:15:46.160 --> 0:15:46.760 in the ground. 0:15:47.400 --> 0:15:49.120 The show Well wow, what. 0:15:49.520 --> 0:15:52.480 Planting a tree is easy? But remember it has to 0:15:52.520 --> 0:15:57.720 happen fifteen billion times. IBM Research has been operating in 0:15:57.800 --> 0:16:01.920 Nairobi since twenty thirteen, and what Kenya wanted, at least 0:16:01.960 --> 0:16:05.880 in the beginning was straightforward. The prith Fee model that 0:16:05.960 --> 0:16:09.200 IBM and NASA built could be used to essentially make 0:16:09.240 --> 0:16:13.560 the world's greatest map, and Kenya, with IBM's help, could 0:16:13.680 --> 0:16:16.720 use that model to make the world's greatest map of Kenya. 0:16:17.760 --> 0:16:20.000 The first step was to lay a grid across the 0:16:20.080 --> 0:16:24.200 topography of the country, break the forest into manageable bite 0:16:24.240 --> 0:16:27.560 sized pieces, each of which could be analyzed separately. 0:16:28.400 --> 0:16:30.680 So, because our forest is massive when you look at 0:16:30.680 --> 0:16:33.360 it in terms of green rite, but only lay it, 0:16:33.600 --> 0:16:36.240 you're able to break it into pieces like into boxes. 0:16:36.480 --> 0:16:40.120 And for us that was important because then it's easy 0:16:40.160 --> 0:16:42.920 to tackle it. When it's in a greed system than 0:16:43.120 --> 0:16:45.600 just as a massive forest. So that was also what 0:16:46.280 --> 0:16:47.680 the model was able to do. 0:16:48.120 --> 0:16:51.680 Then the model painstakingly sorted through each of those boxes 0:16:52.040 --> 0:16:54.360 and look for what Philip calls hotspots. 0:16:55.040 --> 0:16:57.440 So you can see, for example, very quickly, which other 0:16:57.520 --> 0:17:01.160 areas are being eroded very fast, and that you need. 0:17:01.040 --> 0:17:01.840 To quickly protect. 0:17:02.320 --> 0:17:05.120 Yeah, because you sometimes and that's where you want to target, right, 0:17:05.200 --> 0:17:07.320 I mean, it's not possible to do everything at the 0:17:07.320 --> 0:17:07.760 same time. 0:17:07.960 --> 0:17:09.760 Do you have a definition of a hotspot and how 0:17:09.760 --> 0:17:12.480 many hotspots are there according to that definition. 0:17:12.840 --> 0:17:13.480 Oh, there are a lot. 0:17:13.560 --> 0:17:16.840 So we have more than forty water towers, and I'll 0:17:16.880 --> 0:17:19.639 tell you all of them have hot spots. And the 0:17:19.680 --> 0:17:23.200 hot spots in my definition areas that are being degraded 0:17:23.640 --> 0:17:26.600 fast and in a very unusual way. Right, you can 0:17:26.720 --> 0:17:30.320 literally see how human activity is seriously degrading that particular 0:17:30.320 --> 0:17:32.800 area that if you do not have a direct intervention, 0:17:32.960 --> 0:17:36.119 we'll lose the entire forest. So that's the hot spot 0:17:36.200 --> 0:17:38.760 for us, because you think about cutting a hundred trees 0:17:38.800 --> 0:17:40.800 a day and cutting a million trees a day, So 0:17:40.880 --> 0:17:43.800 that's a hot spot. You want to look at places 0:17:43.800 --> 0:17:48.399 where there's just unusually high activity of deforestation. 0:17:48.480 --> 0:17:50.600 In a hot spot. The size of each box in 0:17:50.640 --> 0:17:53.520 the grid was ten by ten meters, about half a 0:17:53.560 --> 0:17:57.439 tennis court. That's how closely they were examining the forest 0:17:58.560 --> 0:18:02.760 so very crudely. The model ingests all of this satellite 0:18:02.840 --> 0:18:06.879 data and it helps you answer some very specific questions 0:18:06.920 --> 0:18:11.560 like where should we prioritize our tree planning efforts, which 0:18:12.000 --> 0:18:16.600 areas down to an extraordinary levels of specificity are eroding 0:18:16.720 --> 0:18:20.560 most quickly. You know, all those kinds of practical questions 0:18:20.600 --> 0:18:22.159 about how to direct your strategy. 0:18:22.480 --> 0:18:24.560 So if you think about a smart forest, right, and 0:18:24.600 --> 0:18:27.200 that's really for us according with smart fencing, smart forests, 0:18:27.200 --> 0:18:30.960 everything that's smart because of AI. If you think about 0:18:31.680 --> 0:18:34.359 your usual what you can see with your eyes and 0:18:34.400 --> 0:18:37.399 then the satellite layer which just zooms in and you 0:18:37.440 --> 0:18:40.320 see green. So what the model has been able to 0:18:40.320 --> 0:18:42.680 do is to create a smart layer, right, and then 0:18:42.720 --> 0:18:46.320 that smart layer you can actually see many things from 0:18:46.320 --> 0:18:49.400 analytics to the greeds to a dashboard on a lot. 0:18:49.760 --> 0:18:53.480 So able to layer to those blocks, you can quantify 0:18:53.600 --> 0:18:57.919 degradation by blocks. You can match intevations, you can match reforestation. 0:18:58.440 --> 0:19:00.679 I asked Philip to imagine what would have been liked 0:19:00.680 --> 0:19:04.720 to attempt the tree planting project in an era before AI. 0:19:05.480 --> 0:19:11.880 His answer was plant fifteen billion trees, restore the water towers. Impossible. 0:19:12.920 --> 0:19:17.080 With Prithvi on Kenya's side, though it's really happening. What 0:19:17.160 --> 0:19:20.000 should be clear by now is how versatile Prithvy can be. 0:19:20.600 --> 0:19:24.159 Want to know how to combat deforestation, Prithvy can model that. 0:19:24.600 --> 0:19:26.280 Want to know when the best time in the year 0:19:26.320 --> 0:19:30.400 to plant your crops is, Prithvy can help predict that too. 0:19:30.800 --> 0:19:35.160 Last year, six months after IBM started helping Kenya with reforestation, 0:19:35.680 --> 0:19:38.840 Kenya needed Prithvy's help on something else and it was 0:19:38.880 --> 0:19:39.680 an emergency. 0:19:40.560 --> 0:19:43.280 So something was happening in the world that we sort 0:19:43.320 --> 0:19:46.040 of had these floods that we didn't expect. 0:19:46.440 --> 0:19:49.040 In the spring of twenty twenty four, Kenya was hit 0:19:49.080 --> 0:19:53.240 with thunderstorms and torrential rain, days and days of. 0:19:53.200 --> 0:19:56.440 It, and so I got a call from the Red Cross, 0:19:56.600 --> 0:20:00.080 the one of my friends, and they're like, Ambassador, we 0:20:00.119 --> 0:20:03.160 need a little bit of help on how we deal 0:20:03.240 --> 0:20:06.240 with response because what we're seeing is unusual, right, because 0:20:06.240 --> 0:20:09.119 normally you would only have one area, all of a sudden, 0:20:09.240 --> 0:20:12.439 we had an entire country flooding. In April we had 0:20:12.480 --> 0:20:16.840 about three thousand, eight hundred kilometers square kind of total 0:20:16.960 --> 0:20:21.280 land flooded, which is unusual for Kenyon. And so when 0:20:21.320 --> 0:20:23.600 I got this call, we were like, okay, there's someone 0:20:23.640 --> 0:20:26.320 could do with IBM. We only did one function for 0:20:26.359 --> 0:20:29.160 the trees. It was actually a climate model, and we said, 0:20:29.200 --> 0:20:34.880 can we use this to help us better respond to floods? 0:20:35.440 --> 0:20:38.520 And so that was how we started having this discussion 0:20:38.560 --> 0:20:42.480 with IBM in terms of repurposing the model to help 0:20:42.720 --> 0:20:46.320 us deal with this new challenge around floods. 0:20:46.880 --> 0:20:52.840 Again, prithvy is versatile. Prithvie could use everything it new 0:20:52.880 --> 0:20:57.240 about the land, the forests and infrastructure to analyze how 0:20:57.320 --> 0:21:01.560 and where and when floods would occur. The Kenyan government 0:21:01.600 --> 0:21:04.000 could then use the model to help the Red Cross 0:21:04.119 --> 0:21:08.440 organize its response, show areas that needed to be evacuated 0:21:08.760 --> 0:21:11.840 or safe places where the Red Cross could set up camps. 0:21:12.520 --> 0:21:15.119 That information was invaluable. 0:21:16.080 --> 0:21:18.640 Historically, what has happened is that they would set up 0:21:18.720 --> 0:21:23.879 camp based on population congregation right where people assembly is 0:21:23.920 --> 0:21:26.240 where they set up a camp, not based on any data, 0:21:26.440 --> 0:21:29.960 right simply because people are there, they will come there 0:21:30.000 --> 0:21:33.600 to provide services and emergency response. What we realize is 0:21:33.600 --> 0:21:36.080 that that model doesn't work. So what we've been able 0:21:36.080 --> 0:21:38.560 to do with IBM is be able to sort of 0:21:38.640 --> 0:21:42.199 give red cause very specific locations or options where to 0:21:42.240 --> 0:21:44.680 set up camps. So if people come here, just tell 0:21:44.720 --> 0:21:48.480 them no, move here, that's the safe place you really 0:21:48.480 --> 0:21:50.040 want to go. So I think for me that was 0:21:50.080 --> 0:21:52.240 really amazing. So we are calling them a very funny 0:21:52.240 --> 0:21:54.480 word for it, flood assembly points. We always have fire 0:21:54.960 --> 0:21:56.879 for assembly points, but now we can say we have 0:21:57.520 --> 0:22:02.280 literally flat assembly points that are safe or citizens. 0:22:02.680 --> 0:22:07.240 That's fascinating. So the model has ingested this incredibly granular 0:22:08.080 --> 0:22:14.679 picture later of the topography and weather patterns of Kenya. 0:22:14.800 --> 0:22:18.399 It's just giving you a set of useful predictions about 0:22:18.400 --> 0:22:20.320 how you should shape your response. 0:22:21.280 --> 0:22:23.840 Yes, and what we did remember is that, as I said, 0:22:23.840 --> 0:22:27.760 it was a full multistate called capability. What IBM gave 0:22:27.840 --> 0:22:30.520 us was a base map. We didn't have that before, 0:22:30.920 --> 0:22:33.480 and a base model. So you cannot have these layers 0:22:33.520 --> 0:22:34.879 up on layers up on layers to be able to 0:22:34.880 --> 0:22:37.000 make intelligent decisions. 0:22:40.720 --> 0:22:43.640 Throughout my reporting on this episode, I've been really impressed 0:22:43.680 --> 0:22:46.720 by what Prithvie can do. But it doesn't stop at 0:22:46.720 --> 0:22:50.520 floods and reforestation. Prithvie has also been used to look 0:22:50.560 --> 0:22:54.800 at wildfires and floods in the UK, and Kevin told 0:22:54.800 --> 0:22:58.040 me that researchers in Africa have even used prithvy to 0:22:58.200 --> 0:23:02.120 identify locust breeding grounds, which could help them prevent swarms 0:23:02.160 --> 0:23:07.479 that destroy crops. But all these are issues on land. 0:23:07.920 --> 0:23:10.119 I mean, I always say to people, seventy percent of 0:23:10.119 --> 0:23:12.000 our land mass is ocean. 0:23:12.520 --> 0:23:15.240 Kate Rice is the director of the heart Tree Center, 0:23:15.600 --> 0:23:20.440 which focuses on adopting AI into UK's public and private sectors, 0:23:20.840 --> 0:23:26.320 and one of those sectors is the blue economy. Oceans, fish, shellfish. 0:23:26.960 --> 0:23:31.720 But oceans are huge, and getting data from motions is difficult. 0:23:31.600 --> 0:23:34.240 So you're dealing with something where there's not a lot 0:23:34.240 --> 0:23:40.000 of people walking around collecting data. So the real difficulty 0:23:40.160 --> 0:23:44.719 is understanding that collecting enough data to make anything makes sense. 0:23:45.320 --> 0:23:51.199 And oceans are very complex in terms of their interaction 0:23:51.760 --> 0:23:55.000 with our climate and how they interact with the climate, 0:23:55.480 --> 0:23:59.520 so understanding the physics space models is pretty challenging too. 0:24:00.200 --> 0:24:05.639 Once again, enter IBM. IBM created a new geospatial model 0:24:05.880 --> 0:24:09.840 to help us better understand our oceans. Heart Tree and IBM, 0:24:10.080 --> 0:24:13.359 along with the Plymouth Marine Laboratory, the UK Science and 0:24:13.440 --> 0:24:17.239