Climate Action 101: Frontier Technologies for Adapting to a Changing Climate | CB Stories
[Music] welcome to capacity building stories climate action 101 a podcast brought to you by the Paris Committee on Capacity Building's Network in partnership with the United Nations University I'm Eva Costas your host and in this series we're breaking down the essentials of climate action 101 style join us as we explore key concepts related to climate change while connecting the dots between capacity building research policy and practical Solutions let's dive in [Music] welcome to Frontier Technology 101 today we're diving into Frontier Technologies what they are how they're shaping climate adaptation and how capacity building plays a critical role in leveraging these Technologies effectively to guide us through this conversation I'm joined by two wonderful guests Himanshu Shekhar and Maxime Souvignet could you please introduce yourselves to our listeners Hi Eva a very good morning and good morning to everyone in tuning in my name is Himanshu Shekhar I'm a senior researcher at United Nations University Institute for environment and human security here in Bonn and I my interest lie mostly in the field of making cities more resilient and adaptable to climate change Maxime uh good morning Eva good morning Himanshu my name is Maxime Souvignet and I lead uh the team for climate risk analytics at UNU-EHS welcome both and thanks so much for joining us uh now Himanshu to start with the basics what is frontier technology and why is it important in bridging gaps and addressing the capacity building needs to adapt to a changing climate yeah indeed the question you know the frontier itself is something that one would say is relative and that's how we can also see the frontier technologies to be something that looks fundamentally different at this point of time becomes a mundane thing and that's what I guess is a basic understanding of frontier technologies these are radical scientific breakthrough that have the potential to basically shift how we do things let's talk about for example when telephone came it changed communication when internet came it changed how we look at things how do we analyze inform ourself similarly there is a new way what we call Industrial Revolution 4.0 and when we talk about uh Frontier Technologies nowadays it's often about the that we hear artificial intelligence blockchain machine learning they have a real potential to change how we interact how we understand and how we inform ourselves so that's what in this sense frontier technology mean it could be seen as and when it comes to climate change particularly adaptation to climate change the unfortunate reality of the world is right in front of us we see the flooding that is happening in Central Europe nowadays or any part of the world we look at indicates to the fact that we need all hands at desk and there is a very essential component that the technology can provide be it in terms of knowledge be it in terms of support or be it simply in terms of capacity that is needed by people who have the knowledge to do it hence the frontier technologies be it artificial intelligence machine learning blockchain and whatnot have a real potential to enhance maybe even transform the way we address climate change that would be my short answer to this question so now Max building on that in the short term what key capacities and resources are needed to seize the opportunities that these Technologies offer yeah I think shortterm is the is the keyword there so as Himan was saying so basically new frontier technologies are similar to what happened with telephone or email etc the difference is and I think the main difference is that it's happening very so the the pace is uh is incredibly fast and in the short term then we see there are already an emerging let's say actors and tools in the in the climate landscape I think from a UN perspective you know these use cases are very often concentrated in countries uh that are developing the tools and I would wish and I think that there would be many opportunity to include social and Indigenous knowledge within the new developments so especially in the global south I think this is where in the short term we would need as as UN entity to concentrate on thank you so now that we have a bit understanding of frontier technology and its potential for climate adaptation looking at data I'm wondering what kind of data is needed for climate adaptation policies uh Himanshu let's sort of take a small step back just to understand you know we understand normally the risk as IPC defines is if we can say it is made of three central components one is the risk hazard itself so for example if there is going to be an extreme amount of rainfall that could be a hazard second is the exposure am I placed I living in a place where it is likely to rain that much if not then I'm not exposed if yes then I'm exposed and third is the vulnerability do I have an umbrella you know or do I don't have an umbrella if I don't for example if I don't have protection against that possible Hazard then I'm vulnerable so this is in this understanding for now we have a better understanding I would not say very comprehensive but we still have a better understanding of the hazards but when it comes to exposure and vulnerability to climate change these are extremely complicated subjects and they need they have thousands of data points for example my vulnerability could be different than yours and so on so how do we have this differential vulnerability into equation that is where I think the potential of artificial intelligence come into play similarly what kind of different exposure is out there that comes into play but also going back to the original understanding of hazard itself even now a lot of understanding comes from the very global and regional level when it comes to the local level there are what we call as data deserts unfortunately still in different parts of the world so the AI has a very strong potential to also provide that sort of data gaps you know uh that that kind of uh information that is needed what we need to move from is that we are often familiar of the concept of forecasting so we know what might happen in couple of days ahead from them from there now we move to go into what is called as nowcasting we need to have real time time and dynamic information at hand and that sort of data is needed so data on exposure data on avability as well as a very nowcasting sort of approach to data it's very essential to adapt to the very extreme nature of events that we see so I also understand that gathering and managing this data comes with difficulty so max what would you say are the main challenges and how can capacity building help address them so I think Himanshu is right with the you know we have to move for we have to add up to forecasting now casting uh it's extremely important extreme when it comes to climate change especially we talk about early warning and so on and so forth I think that there are already a lot of capacities going on but it's and there are also a lot of endeavors also from big tech companies to advance to and then us as a public research for instance in AI flood modeling for nowcasting and protections with publicly available data set that could be used however you know when it when it comes to data then it relies very often these type of models and open-source data to have access to this data they are they are very often promoted in this way the problem is then and that's the next thing that the models they are actually not actually black boxes and that's that's a bit that's a bit of an issue and that's the that's one point I think the other point is you need when it comes to so when you want to make this type of nowcasting or forecasting using new technologies then you need also to have very high resolution data such as remote sensing training in remote sensing is not something that is addressed easily that's very complicated I think but what the other thing what we haven't touched upon yet but and the other type of data that is needed is not just on the let's say the weather the climate side it's also on the impact side so you would need also a lot of data when it comes to damages you know buildings people that are affected and so on and so forth livelihood this is completely missing I think in my opinion when it comes to training capacity building that we can offer people then let's first one on understanding what type of data is needed second understanding how to you know you will not be able to provide capacity development capacity building for remote sensing or it's very hard so that's universities already doing this but I think on the gathering on this damage side data this is there is a huge gap and in terms of what is needed but also what is offered in terms of capacity building in my opinion this is where the biggest opportunity lies for and in the global south related to data and just to go back to a point that you had spoken about Himanshu before so you were talking about how artificial intelligence can essentially help with gathering this data on exposal vulnerability and nowcasting so I'm wondering if you could tell us a little bit more about the role of artificial challenges in this context of addressing the challenges that Max just mentioned and filling these capacity building gaps indeed I think Max has really set the ground for what I was about to address is that we need to move from this hazard based understanding to more impact based understanding so there is more personal what it means is that while it might rain a lot it would affect me differently so the impact based understanding provides this human angle to it it provides a more individual angle to it and that's there that's where the challenge lies not only to gather data but also to interpret it I mean if we have this massive amount of data coming there are very limited capacity to analyze and make sense of it so the role of AI can be across scale at different steps of what how do we gather it how do we process it and most importantly can we also automate certain parts of this decision making I mean it sounds amazing to have this much of data but it also means that we need capacity to distinguish between what is critical what is non-critical and so on so very simply I would say the first part is regarding the data projections itself it can allow us to project a complicated set of data compared to what we do if I am able to humanly have 10 different variables AI can allow me to maybe have even up to you know 10,000 different variables and still do the same sort of forecasting even in a shorter time than I would be able to the linkages the neural networks that are being created that are just mindblowning so the so our role there we need capacity to make sense of that information so that is very important to keep in mind second part is the data analysis exactly so a lot of this data for example what we get for example from our phones we all have become data sources ourselves and we get these amazing you know driving maps navigation apps that we have that tell us about the traffic it's just a compilation of the information that is coming out so we need this sort of capacity also in the climate domain where a lot of information that is collected we need to have clear understanding of what qualifies as something urgent something long-term something medium-term and based on that have automatic understanding or information being shared and then third and I think maybe most important from my point of view would be filling the data void there are parts of the world where still have very rudimentary if in certain cases not at all any data so while it would take a lot of time and resources and rightly so to gather data firsthand we do have tools already at disposal that can allow us to look at what are the characteristics let's say if we talking about a coastal city that does not receive that much of rainfall is surrounded by mountain so do we have a similar set of criteria that fits another city and can we have some sort of data for casting or based on that that model so this sort of what we call data extrapolation can already allow us to fill some of the data voids so my wish for the role of artificial intelligence in this sense is to fill the data voids help us analyze the data and help us with the automation of some of the decision makings I mean of course there is a lot of it's a very loaded point and there is a lot of parameters we need to set we need to collectively decide thank you for that and I like himanshu how earlier on you have mentioned the need for shift from this kind of Hazard based understanding to impact based understanding and really kind of look at the individual and human angle and that actually brings me to my next question because while the two of you have been able to tell us about the potential of transformation of these frontier technologies and artificial intelligence just kind of see what their impact really is it's best to understand a few examples of how these frontier techologies are being applied on the ground so Max Could you tell us a couple of examples where these frontiers technology have been used successfully to advance climate adaptation in different contexts oh yes gladly so I think what I can do is and I can pick up one we we talked about nowcasting basically early warning and then how to deal with these type of events and then as you know as a followup of so that would be one example and then the other one maybe more on the the other type of data such as how do you account for damages after an event how do you get this type of data maybe also using this new frontier technology so on the first example there there's a platform which is called alert California and it's dealing with nowcasting of wildfires it's basically a combination of a lot of different new technologies Leidos for instance so basically very high resolutions camera that are looking at forest combined with remote sensing combined with fire wildfire modeling and all that in real time using AI in cooperation with the private and public sectors in the US then they are able to basically use a network of state-of-the-art technology to support data driven decision in real time that's very impressive and then it's very efficient at basically at not only predicting but and really reacting really fast to wildfire that's one example but there are others such as another tool which is called X view 2 which is it's basically answering one of the main bottleneck when it comes to post disaster so when a hurricane strike or let's say earthquake or you have tropical cyclones etc it doesn't have to be always climate related but then you need to assess what is the damage because then the next question is for instance how much finance is needed you know to to be to be channeled to these countries etc and then x view to is using a combination of several remote sensing satellites with on ground data combined them using AI technology with different type of imagery and then identify damage buildings for instance in it used that in Seria for earthquake that's not climate related but then you get the idea of what it can do I think these are two nice examples that are I think on purpose chosen in also not the global south because I want to make a point on that maybe later on as well great thank you for sharing those stories with us just do flag as well to the listeners if you'd like to learn more about them we'll add them onto the description box of the podcast but you were just talking about the global south Max and especially the application of these tools at the local level with places that actually have limited resources what would you say are the dos and don'ts for leveraging Frontier Technologies and avoiding mal practices yes that's the dos and don'ts that's always a I mean I'm not sure there is a recipe it's also very early stage and it's moving very fast I think I won't go in there but maybe look at maybe some best practices and then maybe some some pathways that we can that we can choose I think first so we talked a lot about frontiers technology AI etc but it's a bit of I mean it's a very large landscape so there is no AI so there are a lot of different type of AI you have you can divide that very simplistically in two big families so you have big AI so it's big tech this is what you know as you know this large language models like ChatGPT that are extensively trained with and very large computer power etc and then you have on the other side the others spectum other side of the spectrum what we call small AI or fringe AI these are smaller tools that are based on the same technology that are also trained on data but then that need you know less data intensive and so on I think this is what we need the direction we need two things so basically this you know small AI is actually very powerful when it comes to local cutting it could be locally customized and it be run on in a local computer's workstations etc you don't need like a super computer to do that I think there're a lot of potentials for the global south countries but also for smaller institutions and so on so that's the I think in there I think there are a lot of entry points when I was talking about indigenous knowledge and social knowledge to include that and to increase ownership using basically this fringe AI where you would have in a way the global south not being the consumer of some AI tools developed and for instance like California by US San Diego the University of San Diego then would be really the developers of approaches that and they would own the also technological infrastructure so that is feasible for this Fringe AI and I think that's a big policy challenge because then I think then you would need to set up then transformational structures and in order to do that but then you have already initiatives that are doing this going in this way that's climate change AI and also every climate AI that are trying to or helping to make this happen and as you suggested then I will post all the links I'm not posting the links but then we will put the links and attach them to this podcast Thanks Max and that's the beauty of podcast where it's all in the future so the links will make it there eventually but I'm really glad that towards the end you touch on this people centered approach and really integrating local traditional and Indigenous people's knowledge into kind of the developing of these froners to technology so on that on this people centered approach I was wondering himanshu who would you say are the key actors or groups at the local level that we must involved in this process and could you also specifically address how we can ensure that the most vulnerable groups also have access and the capacities to leverage these opportunities indeed I think what Maxim was talking about was very fascinating and that rightly ask a question of who gets a say in this I personally view the whole debate about personal frontier technology is not just of the technology or even climate change but of social justice as well we need to have a very strong Justice lens Justice hat and the tools with us to ensure that it does not become another divider hence we need to have all the right actors all the actors I would say on board with us in the process and that's the thing I mean we don't need everyone to do the same thing we need to start with the acknowledgement that a lot of frontier technologies are extremely complex and while we need not have the equal capacity to create it that would be simply futile but we need to have right say at the right point of what are the parameters what are the uses what are the boundary conditions what is the regulative mechanism so while there is a usual list of you know actors that you would like to have we need to look at this debate from the question of centrality and decentrality a lot of this technology is being developed because simply they are so resource intensive by big tech companies so of course there is a bigger role for them there is a role for the government in ensuring that it's just it's Equitable but you also need to think about when we talk about the local level is the role of the academia the role of the vulnerable peoples with alongside comes the role of Civil Society organizations self-help groups the NGOs and the youth what we need to have is provide them first the capabilities to be a very meaningful voice in the debate so there is a capability side of it and there is a capacity side of it so we need to have strong capacities built or augmented at the local level particularly of this group of factors and that's I think that's where academia also comes into play so that they have a very meaningful say in the process otherwise if we leave them behind and that's the second part of the having the capacity it means not only having access to tools but also for example having training in using these tools and that's the beauty of for example what a lot of these large language models what Maxime was talking about they have almost started bridging the language barrier so you need not know English all the time to make sense of the technology technology does that so there is this strong vacular part of the whole debate we need to ensure that everyone has the right tools capacities this is done in a language that they are comfortable participating in and we have to keep in mind that people are solution providers we people are not just data points and the moment we have this shift in thinking because we everyone is together in you know we have to be together in addressing climate change so by keep treating ourselves people everyone as solution provider and then having this data mechanism or let's say digital platform that allows everyone to express their opinion have a part of this decision making automated and then make the most of it I think that can rightly allow us to have what we call a true sense of participation in decision making thank you so much for that indeed I I really kind of love how we talk about this whole of society approach that's really needed and really kind of seeing people as solution providers and now just towards the end of your intervention you brought in how that kind of trickles down into the decision making space and especially when it comes to really scaling these technologies we naturally need to also involve decision makers as part of this process so what would you say are two key areas of support that you would like to advocate for receiving urgent attention from policy makers picking two is always very tricky but I know that we have limited resources so it's always a question of priority I mean now coming from our experience in the UN system as UN University two things I would like to propose first is the ensuring there is policy policy coherence when it comes to adaptation and these policy documents for example a lot of countries are either preparing or revising their National Adaptation Plans they do embed this possibility of digitalization this possibility of frontier technologies unless there is a legal sanctity to these tools there is a there's an architecture for these tools to be meaningfully employed we might simply miss the time and then 5 years down the line we would realize the documents are not providing us enough ground to work so for me embedding frontier technologies in the adaptation and policy-making process to ensure policy coherence is essential in ensuring that our documents are future-oriented they are forward looking because this tool is extremely Dynamic itself so we don't know what how it would look in couple of years time so that policy coherence with some margin for flexibility would be great and second is the capacity of personnel at the local level particularly at the very in the smaller cities in different parts of the world we need to ensure that people are not only familiar with how do you use your computer so there is you know this whole step of digitization digitalization and digital transformation so you not only have Maps digitized but you also have information from this map that could be read through a software and then finally you have the set of information for making that decision and your process changes but this whole set of capacity building training that needs to be provided I feel that is very essential without that even if we have best of the technologies we still would not be able to make most of it that really goes to show that capacity building at the end of the day is a an enabler of of all these Innovations but then also as well in terms of really empowering people to actually like you just mentioned before become the solution providers that they are so you mentioned as well this kind of need to kind of look into the future and this kind of where we looking approach and I would like just to take a step back and actually look at the bigger picture of where this is all heading so Max how do you envision these frontier technologies influencing climate adaptation in the near future yeah thank you very much I was very much thinking about what Himanshu was saying it's not because you have a tool that the job is done and I think this is also looking at the bigger picture in the context of climate adaptation and capacity building it's also not because you learn how to use your tool that the job is done I think you know we have nowadays Tech developers big Tech but also for Fringe etc they have the tools this is existing but then sometimes they really lack the necessary perspective so or what needs to be addressed or they don't have access to data or not to resources to really then address the use cases that need to be addresses so I really think that there is a need for a tool there is a need for capacity building but there is also a need for purpose what is needed so perspective and that's also the role of climate policy making I think then what we need at the end what is required is matchmaking matchmaking between AI Innovations and also collaboration between private public institutions but also and Imanshu was mentioning that Academia research so all that needs this matchmaking between also Global South and or especially within the Global South I think that you need that we need some really forums that some of them are already existing for instance the the AI Alliance it's an an initiative to to enable Partnerships in Global Soft countries with international agencies local communities and so on and so forth and then they advocate for climate data ecosystem that would fit within basically the and enable AI for climate I think there are also in a way and I was talking about that before emerging Capacities especially for edge AI in the global South and I think these capacities that are emerging right now they need to be strengthened with data access infrastructure local control of under Tools and all that I think that would be you know it would lead it would influence greatly and dearly climate adaptation very soon in the coming years I think now frontier technologies are also advancing and evolving really fast maybe even faster than our ability to keep up with them so I'm wondering himanshu what would you say are the unknowns and where do you wish to see additional support in the coming years that's I think a very deep question and I think maybe this I would take this more as a que to be cautious about the whole thing it's an amazing tool there is no denial about it and there is a lot that we can do but it's so swiftly I don't know if even swift is the right word it's such an exponential growth in the technology itself we just don't know how it would look like in five years down the line or even two years down the line so how do you regulate these things how do you provide so that's for me is a big question and I think a lot of you entities are working on answering this question second and maybe even more concerning part is what we call the rebound effect we don't know the environmental impact of these emerging technologies themselves at the same time the societal impacts it might end up becoming another tool of inequality so these are the things that I would be more cautious about but at the same time where I wish to have additional support is of course you know being caucious of these things I if I have to pick maybe two things again one would be that we need mass capacity already of the key actors we may not be able to do it for everyone right away but at least the key actors that are needed they need to have the right capacities and training so that not only they understand what the tool is about but the right perspective what Maxime was talking about you know having a tool but knowing what it is capable of is equally essential and second I would go back to the policy making process again there are dedicated team in almost all the entities all the countries that take part in the policymaking process and they need to have a very good understanding of the tool they need to have very strong capacities to understand what implication it has for policy making be it for National adaptation plan National communication NDCs and so on so those who are preparing these documents they need to be prioritized for capacity building thank you himanshu and and bringing it back to you Max as a final thought how do you foresee the pragmatic applications at frontier techology and advancing adaptation for climate change in the next years yeah that's also a good question so it's a pity it's recorded because then we can come back to this podcast and see how wrong I was in my predictions but and still I will try that and but I think I'd like to to remain a bit more a bit more General and then bit summarizing a bit what we have said that I think to harness the potential of for climate action then the Innovations coming from AI Frontier Technologies in general they need to bridge between global divides and it's really important that it you know we foster collaboration between disciplines so that's not just qualitative approach needs also to be you know not only quantitative needs to be qualitative also include social knowledge it's very important it's to bridge across regions communities and we need to ensure that these new technologies they are not only enhance the ability we have as country as a people etc to combat climate change but also we have to ensure also as organization that they do that in a way that is fair that is in a way that is transparent and also know rooted in the knowledge in the countries where it is applied in local knowledge I think that's very important and it's just when we basically will manage to do that we will be uniting different type of technology different needs different perspectives that we will use the full potential of new technology to address climate change and climate adaptation in particular thank you so much I think on that note again I love how you just towards the end of your intervention you show that indeed back to the importance of this country ownership but also people ownership and these people center and participatory approaches through these Technologies but then I also goes to both of you your interventions through the episode as well show that the more these technologies advance as well they come with a big moral and ethical responsibility from our side naturally and then also for other international organizations so again I just would like to thank you both immensely please be sure to check out other episodes under the CB stories Climate Action 101 series and thank you for listening [Music]
2024-10-29 18:43
