Climate Week NYC 2023: Harnessing the positive potential of AI for urban climate action

The New School: Good morning, everybody. We're gonna get started. So if you wouldn't mind taking your seats, I am really pleased to be here. My name is Renee White. I am Provost and Executive Vice President for academic affairs here at the new school, and it really is a pleasure to have the opportunity to welcome you. So we are really glad that you are here to do the work that's so important, and to participate in this morning's event. I do also want to make sure to thank our master of ceremonies, Josh Soren. So thank you. To our distinguished panelists and to all of my colleagues here today who have worked to organize this important presentation and discussion, especially Professor Timon Mcpherson and associate Director Christopher Kennedy, from the Urban Systems Lab, as well as our partners at Google Org. the Center for public Impact and World Resources Institute. The New School: We have all directly experienced in various ways the impacts of climate change which are increasingly real and visible in our daily lives. This is especially true. In cities like New York.

we're wrestling with extreme heat and rainfall. Smoke from wildfires, sea level rise and coastal flooding simultaneously. Technologies like AI are rapidly transforming. How we interact with the world around us in profound ways. amidst concerns about its power. Is there room for great hope in AI as a tool for protecting cities and communities from climate disaster, and also addressing the ongoing climate crisis? So to address this question. We're excited today to convene senior leaders from across public. private and social sectors to discuss showcased AI solutions that could very well reimagine the immediate future of climate resilience in cities, and, I just want to add, I had a conversation yesterday with a friend and colleague of mine who does a lot of different kinds of work and publishing, and he was actually talking about the ways in which so many different industries have to harness their shared talent and knowledge and ability to communicate and reimagine what solutions could be, and to really provide spaces of opportunity, not only for us, but obviously for our future generations. So this morning we will welcome demonstrations of 2 novel AI climate, resilience, solutions. climate, IQ. Presented by the new school's urban systems, lab

The New School: and data for cool cities presented by the World Resources Institute, and on behalf of the new school, I would like to take this opportunity to provide our thanks to Google org for their generous award of 5 million dollars to advance the urban system. Lab's climate, IQ. Effort, which we hope will spur continued research and faculty student engagement around the critical issues of climate change and emergent technologies. So, following these demonstrations, we will have the opportunity to listen in on perspectives from our impressive panelists on the implications of governance and technology for climate resilience. The New School: This panel will be moderated by Joel Towers, professor of Architecture and sustainable design in the school of constructed environments at Parson School of Design, as well as a new school university professor, and also Co. Director of the Tishman Environment and Design Center. What we call Tedsee.

We then hope you will join us after the panel for a networking event, with coffee and light fair just outside the auditorium. So in close. It is one to say that the new school is deeply committed to collaborative and interdisciplinary research and scholarship, to address climate and environmental justice issues from the work of our centers, including the urban Systems Lab and Tishman management at the Milano school of policy management and environment, in the schools of public engagement The New School: to our undergraduate program in environmental studies housed in both the schools of public engagement, and also Eugene Lange College of Liberal Arts. to the work being done on sustainable fashion practices at Parsons, and as provost, I am so proud, and and feel so lucky to get to be in a space where all of this incredible work is happening, and I'm inspired by the faculty, the staff, the students who are so clearly, deeply committed to really advancing justice principles, and to really come up with solutions that are not just. The New School: but are also sustainable, and that can be replicated so collectively. The new school and its colleges are proud to challenge convention through courage. creativity, and cooperation. The New School: Projects and conversations like this are central to our mission of generating positive change for the future. So it is our pleasure, and it really is my pleasure to have you all here to participate in this discussion. So thank you again for joining us, and I'm going to pass it off now to Josh Soren. So thank you.

Morning, everyone, and let's give another round of applause to Dr. White for those great opening remarks. The New School: So my name is Josh Soren. I work for the center for public impact. We're a global nonprofit that was founded by the Boston Consulting Group. And we're one of the partners that are thrilled to host this event, and just want to thank you all for for joining today The New School: to date. Cpi has worked with over 200 local governments across the world with the mission to improve their capacity to solve complex problems. In my role, I lead our global portfolio of climate work across the world with a goal of increasing the speed and scale of climate action

The New School: across the world. Our organization was founded on this understanding that the challenges that governments face today in the 20 first century a very different nature than the challenges governments face in the century when their systems, structures, and processes were set up for. The New School: and so to address these complex challenges. especially those associated with climate, mitigation, adaptation. The New School: we think about, how can we increase the speed and scale at which they're learning and improving on the on the work that they're doing. So, not just following the traditional process of analyzing, planning, and delivering, but really viewing experimentation and learning as a core way of core part of the the work that they do. The New School: This is why I'm so excited for today's event, and also why Cpi was honored to support the Google Org Impact Challenge on climate innovation which committed 30 million dollars to fund big bet projects that accelerate technological advances in climate information in action.

The New School: Just a minute. We're going to kick things off with some opening remarks from Dr. Mahmoud Mohildin, the UN. Climate change, high level champion for Cop 27, and also the UN. Special envoy on financing for the 2,030 development agenda, as Dr. White said, following that we're going to see 2 showcase presentations of of solutions that were 2 of the winners from the Google Google Org Impact Challenge on climate innovation. The New School: The first will be from the new school's urban systems lab, and the second will be from Wri. Following these showcase presentations, we'll have a panel of experts who will discuss the opportunities and challenges The New School: of leveraging technologies like artificial intelligence, to accelerate climate action.

The New School: Following that panel, we'll also have a Q&A. So please do get your questions ready throughout the the panel. The New School: And then, as Dr. White said, following that, we'll have some networking afterwards, we'll get to connect with each other. The New School: So with that I'm very pleased to hand it over to Dr. Mahmoud Mokiyildin to give us some opening remarks. The New School: Thank you. Please help me in welcoming them to the stage. The New School: Well, thank you so much for the kind introduction and the opportunity to discuss with you issues related to climate change, sustainable development, and raise more questions, perhaps, than giving hints The New School: to answer. So I'll be speaking on AI Hs and Nap The New School: to be explained.

So AI. The artificial intelligence and its implications. The New School: HS. Is human stupidity that resulted in the mess that we're in in climate and sustainable development goals. The New School: And Nep is my own acronym for acronyms, please, especially if you are addressing wide, broad audience, who are not familiar necessarily with the jargon that you are using. The New School: because I feel, after getting exposed to some of the recent discussions about artificial intelligence, their policies, regulations. That's not just a matter for the top professionals working in their labs and their silos at the issue. But it's basically an issue of concern for everybody now on the planet.

And before I proceed, I'd like to acknowledge the the sponsors and the the partners in this important work, including Google Org, which I started my week in Europe by attending many of their sessions. The New School: focusing on localized solutions based on adaptation and resilience in many developing economies, including from South Asia and and Africa. I'd like to recognize as well the work of Wri, and have been very much involved with us as partners in climate change and assessment of progress or lack of progress. The New School: Cpi. It had been already introduced with its good impact in our work, and of course, the new school, and this is my first time in person to be in the new school. But I have many good friends as professors. old researchers, and graduates from the school. The New School: As I was sharing with one of some of your professors before this session.

The New School: I graduated from a university called Warwick University in the UK. And this is a university that is known to be Mrs. Margaret Thatcher's favorite university. So you can't really expect the kind of interesting discussions and debate from a graduate of Warwick and a graduate of new school. But this kind of free change The New School: of different views makes matters more exciting than when you talk to people who are from the same school or same ideology. But let me just outline to you the kind of landscape that we are trying to discover, and where exactly issues related to settlements. The New School: Come into that bigger context of matters.

The New School: We had a very big fight for almost 2 years to prove to everybody who was participating in what so called climate action, trying to apply the Paris agreement in these different areas of work in mitigation, adaptation, resilience. Now, some decent work is happening in loss and damage. And of course, finance. The New School: that climate action is part of the sustainable development goals, not just by the definition and the construction of the sustainable development goals, but basically in every kind of work that we are trying to do in the Paris agreement and its different components in mitigation or adaptation. You are touching, positively or negatively something related to the Sdgs The New School: energy and energy efficiency and access to energy and dealing with renewables is one of the core goals and targets of the Sustainable Development Goals, which is Sdg. 7. The New School: Adaptation which you are discussing today a part of the adaptation agenda had been neglected for many years with a false and generous assumption that the world is going to be doing a fantastic job in mitigation and reducing the emissions.

The New School: and that was neglected. It was bypassed even in the discussions, until I would say, during the Shaman Sheikh The New School: discussions coming from a cop that is being very much known as the African cop. At the same time The New School: issues related to adaptation were put really in the bigger context of matter. We have the Shamachi adaptation agenda that covers areas related to food systems, agriculture, water management settlements, including urban and rural developments. In addition to that, there are issues related to deforestation and dealing with the threats The New School: related to to forest and coastal areas. So this is the adaptation agenda at large that requires finance, technology and behavioral change to to support. Now, when it comes to the Sdgs at large, the Sustainable Development Goals, the 4 human made reasons and shocks

The New School: and neglect a lack of good prioritization. The New School: We are only on track of 12% of the 169 targets of the 17 team goals The New School: we managed to measure more or less 140 targets. Of the 169 were only on target of 12%. 50% of the targets are slightly or significantly off track. The New School: and around 35% or more slightly more. They are worse than where we were back in 2,015, The New School: those the politicians that we met, many of them around the like to blame it on the war in Ukraine or on Covid and the shocks associated with the health crisis. But actually we were not on track even before 2,019, 2,020, and before Ukraine, so we cannot really blame it on these external shocks. But basically it's because of lack of progress from the very beginning, on not taking matters related to investment in infrastructure investments in human capital which are more important and investments in resilience. And we are now

quoting the editor from the Financial Times, who moderated the session and the UN. A couple of days ago, saying, Well, it's like in football soccer. We finished the first half The New School: in a very bad way, and we have the second half now towards 2,030. If we are going to be continuing doing that, we're going to be missing the goals and targets with more suffering because these are not just numbers. These are basically issues related to people are getting more poor inequality, deterioration in health systems, education compromise of our infrastructure and further deterioration in climate and The New School: diversity. Sometimes statistics hide behind them what we are concerned. So there are lessons from the old Millennium Development Goals.

The New School: which were applied from 2,000 to 2,015, and the successful countries like China, India, from middle income countries like Indonesia and some middle income countries in Africa, Latin America, we managed to do better than others in the Millennium Development Goals. The New School: When we have better data systems, better finance and better implementation with good institutions and policy framework. The New School: The Sdgs should benefit from this experience, but seem that we're not really benefiting much from the goods and bads of the Mdgs. The New School: Now, with the with the Sdgs we, we need more of the same data finance with implementation, and as far as we can really have The New School: artificial intelligence. The New School: providing positive augmentation of all of that better data systems for decision making and harnessing technology for good or for the good and doing better in understanding behavioral change and influencing behavioral change. The New School: So what we need then, in specific areas, of course, the experts are in the room. But I would expect that matters related to better settlements and better construction and better designs, and we have architectures as well. Urban planners in the room.

The New School: We have first the issues related to the low carbon economies. This is meaning for us, the 0 targets doing better in energy and doing better in transport and transportation between energy transportation and decarbonisation. If we can do a better job, whatever you are going to be doing in resilience and adaptation will be much better on energy. You can think of utilization of AI or artificial intelligence. The New School: to minimize emissions, to have better management of the new renewable energies, and they suffer from the fact that they are intermittent, and they are very much dependent on expensive resources and batteries. So if AI could be utilized into that direction, that would be a great plus on transport, better traffic system The New School: with digitalization of this traffic system. And I'm happy through that our work issues related to the digitalization of the traffic systems and relying more on electric buses in Africa and South Asia, and some of the small islands have been among the winners of the pipeline of bankable project that we are working on with the UN. System BC. And many others on decarbonization, especially in the hard toads. Sectors like cement fertilizers, aluminum, and steel, especially when it comes to cement, aluminum and steel, as inputs in the in the construction and building material. A great deal of work is required here in order to utilize artificial intelligence, to guide us on the priorities. But in all cases we'll need finance The New School: on the shamsheet, adaptation, agenda, food, water again, settlements, agriculture and dealing with the requirements, I think, coming from a farming background, I would appreciate really more guidance in areas related to protection of rural dwellings.

The New School: People tend, especially in academic institutions. Top ones say there are urban areas and rural areas. Fact of the matter is like my village is not urban. It's not rural. It's like some of my former World bank colleagues call it urban. So you have features of rural features of urban, and perhaps you are losing many resources because of that. But when it comes to agriculture. The New School: inefficient irrigation, research, and farming would be very much super helpful to the work. And again, in issues related to productivity, and how to enhance the contribution of the farming communities. That would be a plus having said that. And I'd like to leave you here with some of the thoughts, thoughts, and great issues of concern, many of us, especially of what so called the global South. The New School: We read about artificial intelligence. Many of the people from the global South study in top notch universities and elite universities, including the new school. And there is that issue of fear of of technology. And there is the the concerns about disruption that I wrote a brief book with colleagues on business governments and Sdgs with a focus on disruptions of technology and how

The New School: smart disruptions could be could be helpful. I can see that many people are trying to regulate and super regulate AI. And I think the history of dealing with technological advances is basically about identifying the possible good, the possible bad, and trying to maximize the goods and deal with the concerns. Yes, there are differently, as all technologies have informed us, there could be implications of resources The New School: and artificial intelligence would be dependent on energy more than necessary. Can we be more efficient than that, or in core rare material The New School: that could be the subject of battles and wars. In countries in Africa, especially with the unhealthy competition between some of the big incumbent powers like America or the US. United States, or the emerging powers like China and India. So here can we have really some establishment of the rules of the game when it comes to access to core rare material in areas related as well to inequality. And this is the old argument forever. When you have technology there are losers. There are gainers. There are some people, and I always read The New School: writing an introduction to the oldest book in Arabic on Economics, and goes back to 1,908, of course, came late after Mr. Adam Smith. But anyway, in that introduction of the book issues related to challenges of technology, and how President Santana of Mexico complicated the introduction of railways to protect the the service of transport that was conducted mainly through The New School: donkeys and donkey keepers. The good motive there is to keep an old industry from competition, but he delayed his country for years. We know how China, for instance, missed it when they were putting barriers on trade and investment for many years, and that's why they were behind. Since 1,820. We know attempts.

including burning the the chair of wisdom of Mr. Jakart, who developed machines to make us more efficient in producing clothing and people who were basically doing it manually felt the the threat of the machine. So typical arguments about Luddite's kind of rejection of new technologies need to be harnessed by The New School: good understanding. The New School: good policy and good regulations and good interactions between the top thinkers, between the policymakers and those who could be benefiting from that our concerns in the global South, that while the advanced technologies are captured and protected, there could be some sort of barriers to access. And then we have the issues as well, of regulations that could deny access to more information and better and better knowledge. For us. The New School: Leapfrogging will be exactly right, and whatever fears that could come from artificial intelligence, they wouldn't be as bad at the harm that was caused for many years because of human stupidity in wars. The New School: lack of interest in doing good in investments in human capital, including health and education and lack of good prioritization when it comes to policies. Having said that, as I said, I have many questions to you, especially in the policy of the regulations on how best can we integrate society and get that

The New School: kind of benefit from AI, especially when it is close to the people much better than just saying, well, we discovered AI and its cousins or subsectors like machine learning, and we are going to regulate them. And we're going to be putting some codes. This is not helpful. We believe more on free competition. The New School: onto stability and access to information, knowledge, and education, and then the good dynamics will prove themselves right at the end not to outsmart technologies or ourselves. Having said that very grateful, I'll leave a copy of this book for your consideration at the library. If you accept this small gift, and many thanks and good luck. Thank you. So let's give another round of applause to Dr. Mohieldin. We're going to be holding a raffle for this book, and The New School: thank you so much for that. For those inspiring marks. We're now gonna move to the showcase part of this session where we're gonna be hearing from both the new schools urban systems lab. And then, followed by that Wri World Resources Institute about 2 solutions that are AI enabled that can help cities build climate resiliency. And, you know, having gotten a close look at some of these solutions. I think they're incredibly exciting The New School: again. After that we will move into the panel discussion. So with that, please join me in welcoming Dr. Timon Mcpherson to the stage to do his presentation. Thank you. The New School: What's not just about cities? That's for sure. We're going to focus on that a little bit. Thank you, Josh, thank you for those great remarks to really set the scene for us. And good morning.

The New School: I realize the presentation is not loaded yet one moment. just on that. The New School: Remember, when New York City looked like this. this was just 3 months ago. The New School: when New York City had the worst air quality in the world.

The New School: or remember when it also looked and felt like this. This was Ida, which happened just 2 years ago. The worst rainfall driven event flooding here in New York City in our lifetimes, which actually broke a record that was set just a few days before for extreme rainfall. The New School: I mention this because the way most people experience climate change is through climate fueled extreme weather events. and these events are becoming more frequent. The New School: They're not only becoming more frequent, they're becoming more intense and they're lasting longer, which makes heat waves, droughts. The New School: coastal storms, floods. The New School: winter extremes, even much more impactful. It means they're impacting more people. And they're costing more. Here in the United States. We've already experienced 23 billion dollar disasters in just this year alone.

The New School: So on top of this, we also have to adapt to a new climate, normal of individual extreme events, because we have to prepare for events. That co-occur give you an example. Extreme heat and humidity after hurricane. Maria devastated the infrastructure in Puerto Rico in 2,017, and then caused a health crisis. The New School: But let's look beyond these events. What else do we know? The New School: Last year, with hundreds of scientists around the world, we released the outcome of a 7 year assessment effort to assess the global state of climate change by the Ipcc. I'm sure you've seen a lot of this. I want to give you 4 quick takeaways from this report

The New School: the first one. And we've been talking about this already is that climate change multiplies challenges that we're already facing, including food, energy and water security. And basically it makes them worse. The New School: Second. The New School: it's the poorest and the most marginalized who are facing the worst impacts. And thus they're the ones who really need the most support, including financial, technological, and institutional support. The New School: Third, and to me this is really the most sobering one that came out of this entire report is that the next 20 to 30 years of climate change are already built into the climate system.

The New School: So let's put this in perspective. What this means is that about 16 times as many people may face extreme heat as are already facing it now. and you're probably very familiar with the simmering hot summer that we've just gone through. The New School: It also means that up to 1 billion people living in low-lying areas may be exposed to coastal flooding. So to me, a way to kind of interpret this, this data is to say that no matter how aggressively we focus on reducing carbon emissions which is fundamental, it's even existential. We're not likely to stop the acceleration of climate impacts on our communities for multiple decades. The New School: So we need to be focused, as we heard earlier on adapting to climate change as much as we are focused on halting climate change. So here's a question for you. The New School: how much of global finance do you think is focused on adaptation? The New School: Any ideas? The New School: 10%. 10%. Right? The New School: Clearly, we have to aggressively scale up finance investments in adaptation and resilience, including through engineered approaches, through nature-based approaches, also social and institutional transformations.

The New School: The last point I want to make from the report is that what we also showed is that the impacts on people, infrastructure and economies will be largest in cities and urbanized regions. And that's because they're concentrated in those areas. The New School: So this also means it makes cities the locus for where scaled up investment and adaptation can have significant, positive impacts on people's lives. In short, cities can be turned from problems into solutions. The New School: But they need open access to the best technology. Even if we solve the finance gap. they need open access to the best technology and data to be able to make adaptation investments and decisions in the places where they can have real impact on people. The New School: Give you some examples. Here. The city of New York has spent millions of dollars on investments in technology on investments in scientific research and data to make informed decisions on where and how to adapt to coastal flooding, to urban flooding to extreme heat.

Here's one I want to just dial in just to show you how we've been doing this, and what this means for cities who don't have the same resources The New School: in 2,018 after Hurricane Harvey devastated Houston, Texas. We realized in New York that we just don't know enough about how extreme rainfall may cause flooding here in our city. The New School: So we set about to change that The New School: in a large team modeling effort that we co-led here out of the new school. The team spent 3 years using the latest hydrological models to generate new data for decision making The New School: in May 2,021, before that Ida storm that I showed you. The city released the Stormwater Resiliency plan and included the outputs of that technical effort to help us learn under different rainfall and coastal sea level rise scenarios where it's going to flood. The New School: how deep it's going to be. And that information was provided with high spatial specificity. The New School: So now we are able to better prioritize millions to billions of dollars in flood resilience investments where they can have the most impact on reducing exposure and risk of flooding for people for their homes, for their schools, or the critical infrastructure that they depend on for everyday life in the city. And this is just one of many examples we could all talk about and probably show from well resourced cities

The New School: and how they're advancing resilience and adaptation through advanced technology and data. And it's built on data like you see here, that's truly impressive. But frankly, most cities, even most communities don't have. And access to this data. The New School: In fact, just outside the city, here. Towns outside New York don't have access to the same data in the US. Alone. There's over 5,000 cities and towns The New School: who have limited access to climate risk data and globally, there's more than a million cities, towns and communities, formal informal areas who lack local scale risk information. So they just basically don't have the data The New School: to ensure that the trillions of dollars of investments in development in the next decade will also protect, prepare, and help them respond to what the increasing climate challenges are coming their way. So the question really is, how can we drive forward a climate resilient development agenda? If we don't know what we have to be resilient to The New School: imagine. If we could bring similar or better data to all communities and across the world to make those decisions so that we can safeguard people. And that's exactly what we want to do. This is the showcase we want to present to you today that we're excited to talk about. We're announcing a major new effort. We're calling climate. IQ to open access to the best available climate risk information, but also to take the next technological leap to leverage AI to learn from and provide access to the rapidly changing data and climate landscape. The New School: I'll just introduce this a little further. Here are some of the goals of this.

The New School: First, we want to create a next-generation AI environment to drive forward latest advances in machine learning and computation and data integration to provide more accurate higher resolution, climate information, especially for multiple hazards, for extreme heat, drought, air, pollution, and flooding. The New School: Importantly. And this, I think, really picks up on the points from the previous talk. The key here is to democratize the access to make this open and available to everyone. The New School: We want this technology to be able to help support cities, towns, and decision makers at all levels in order to inform planning to prioritize adaptation decisions for for impact and also to unlock adaptation finance in the areas that need it most. The New School: Who's it for it's for everybody, but especially as we've been talking about communities, towns and city level decision makers who really need this hazard exposure data so they can pinpoint the places where they need to focus what are likely going to be limited amounts of funds that have to be as impactful as they can be.

The New School: How can you use it? Well, outputs, as you can see here from climate? IQ will be fed into an open access public dashboard, enabling users to view current near term and future risk. The New School: and to do that for multiple types of hazards. probably wondering, since this is pretty ambitious, how this is actually going to work. So let me walk you through it just a little bit. Briefly. The New School: We brought together an incredible talented team across the world, including support from Google engineers to help us develop the AI infrastructure and the data back ends. We're going to be leveraging machine learning through techniques like convolutional neural networks. We're going to leverage lots of sources of big data and also multiple climate hazard modeling environments. So this means bringing together in some ways for the first time, outputs from hydrological models, from weather forecasting models.

The New School: from climate projection models along with data on land use buildings, roads, and other infrastructure that are all fed into the machine learning environment so that it can continuously learn and improve its ability to predict exposure to multiple climate hazards. The New School: On top of that, we've got to test it. So that means also validating and verifying those outputs with other kinds of event databases, but also critically, together with partner cities. The New School: This is my last slide. We're going to be developing and testing climate. IQ, in a whole range of cities. That's kind of key for the impact here. This is so the AI can learn patterns in large and small city and urbanized areas in formal and informal areas in temperate tropical and desert climates and in older and newer and developing cities. The New School: So to make this happen, of course, requires a lot of partnerships and the ability to look across a range of city types that can feed this critical, local, national, and global data into the AI core that can then produce outputs for any city that has similar characteristics. The New School: We're super thrilled to be able to be partnering with the city of New York as our first city partner to help us develop, test, and validate this approach and thank them for coming on board early with us, giving us a vote of confidence and moving this forward, and we're in active discussions with a number of other city partners. I also don't want to mention that our team includes a whole range of experts, including from climate sense, a climate tech startup, which you saw an early prototype of their Ua Uiu X environment that was developed with the Australian Red Cross and the city of Melbourne.

where they also have amazing partners from the Bayer Institute of Ecological Economics, the Stockholm Resilience Center, the Virginia Climate Center at George Mason University The New School: and the Kerry Institute of Ecosystem Studies. That's just north of here, just outside the city. The New School: Finally, a key component of this is the volunteer team of Google fellows that we'll be working with. And as we heard the critical support from Google Org's climate innovation program that makes this possible. So we thank them for believing in this and seeing the ambition that needs to be solved, to really address adaptation globally. We're also very proud to be able to call the new school home for this project and here at the Urban Systems Lab. So I'd love to tell you more about this, especially afterwards. The New School: We're going to move to the next showcase here in a moment. But I just also want to say that please stay in touch, follow us on social media and subscribe to newsletters to get updates, and if you work for or with the city we really want to hear from you.

The New School: please reach out to us because we'd love to talk with you about how we can work together to initiate what we think is the beginning of a wider moonshot effort to address the global climate adaptation challenge that we face. Thanks so much. I'm also going to invite up my colleague from Dooy Evan. The New School: to tell you about the amazing work that they're doing as well. And so

The New School: thank you, Simon. Thank you to everyone here for joining. It's a busy week, but it's so exciting to get to speak to you. My hope for today is to unpack the term AI. We hear it a lot. You can't read a newspaper without seeing the headlines. You can't throw a stick and hit a conference here at Dunga. It's focused on AI for something, either in concern. So how are we going to regulate AI? How do we prevent harms or in opportunity? And so I want to go a little deeper and unpack. How are we using AI today? What are we actually using it. For what changes when we use AI in our work on urban climate action? And how might that be different in the future, with future systems along the way, I'll touch on work we're doing with Google. I'll touch on some very exciting products that are actually coming out from many others, and we'll try to weave this together into a picture of what we can actually expect from these technologies and what we should really be concerned about The New School: what we should really be hopeful for. So we'll walk through in terms of what AI is helping us do today today. And for the last few years Wri has been at the forefront of using AI to help us see with local resolution at global scale.

The New School: What do I mean by that? Let's take a trip back in history back to 2,014. You're seeing a data set here called the Human Global Sediment layer. It's kind of beautiful in its own celestial way, it looks almost like a constellation of stars. But imagine trying to use that data set to make any sort of urban climate action. The New School: I can't personally think of an action you might be able to say, there's a little bit more settlement over here, or there's a new settlement that's come online that we haven't seen before, but almost inevitably people living in those settlements will be known and registered, they'll be aware. And so you really can't imagine much action, but flash forward a few years, and we've had revolutions in a number of technologies that then combine into this sort of The New School: umbrella technology. In term we call AI. We have new satellite imagery. Sentinel 2, which is a great sensor for these types of issues. 10 meter resolution. We're starting to get to the point where we can actually see urban environments came online in 2,015. We see training data and machine learning and neural networks. We see revolutionary work to make those applicable from folks like Google. And we see platforms like Google Earth engine that allow scientists and researchers to actually do this machine learning for their problems for their communities. And to make this work.

The New School: all of that adds up into some incredible tools that we have today. we can see the challenge. Now, we can see urban heat Islands from satellite imagery, and understand where there might be places where we need to put in place efforts to mitigate extreme heat. We can also understand how human activity at the urban level is changing this. On the right you see a recent paper that we published focused on types of land use and informal settlements and various forms of land use that actually either contribute to solutions or challenges when it comes to people experiencing extreme heat The New School: in addition to just seeing the challenge, we're also starting to be able to track the solution. So you see here our global forest watch product. And we earlier this year, launched a new forest product that allows us to see for the first time at the full global scale across the tropics that Gfw. Works at down to the level of individual trees. So you can actually track if commitments were made for urban greening where they followed through on, and that could be consistent across hundreds and hundreds of cities around the world, allowing us to do research, allowing us to compare, allow us to understand economic benefits and turn that back into better arguments about how we should do better. Urban planning in the future. The New School: All of that adds up to incredible planning. This is nothing without connecting the data and tools that I've described. What AI gives us with work on the ground. You'll hear later on the panel from my colleague Jaya, who's on our Ross Center team in India, and every day they go out and work on plans. They work on strategies, they work on, knocking down barriers to to implementation of these tools with local policymakers. That's where the action is. But all that action can be guided, directed and supported by these AI tools that allow us to see the problem and monitor and track the solution better. The New School: Now, if you take a step back and you look at the most recent headlines that aren't AI satellite imagery doesn't actually come up much when we talk about AI today. Instead, most folks are concerned with a class of AI that focuses on language that focuses on generation of language, these large language models. You've heard of them. Every tech company has one. There's dozens of other tech companies popping up every day with a new model.

The New School: And we're left with this question of How will AI help us tomorrow? We understand that it allows us to see things, to see problems and understand and plan and act. But this set of language technologies, what can actually yield for us as folks who care about urban climate action The New School: in our work. We're finding that where Llms are useful are 2 important challenges that the climate community faces overall. We have a lot of data, but we don't always have insight. We don't always have a clear view of what that data might mean for a specific situation, and we find that Llms can actually help with that insight part of the equation. Similarly, once we have insight, there's often barriers to action. Things take a long time. It can take a long time to do certain small tasks, and then, when all those tasks stack up to a project The New School: or an initiative, or to a mayoral administration, it can take a long time to actually get to action, and while not perfect, and I could share a little bit in the next demo, we believe that Lms. Can actually shorten that timed insight, shorten that time to action and overall allow us to deliver more for our communities.

The New School: Now I should stop here before I go into a demo and say that at Wi we have internal policies around this with our team. These are just our policies. They're relevant for our context, they may not be relevant for yours, but I'll share them hopefully to to kind of make sure that you're seeing this with the appropriate guidelines first is to keep the human in the in the loop. We asked our staff if they're going to use these tools The New School: to actually check the outputs, they are still responsible for the outputs of any work done with Llms. As a person, as an author, of a report, or as someone who's producing a dashboard. You're still responsible. Second, don't share sensitive, confidential, or private information. I think this is a general policy. Even folks like Google, if you look at Bart, it says, don't share anything. You wouldn't want someone else to see here. This isn't because we don't directly trust these tools because they're under development. It's new technology. We need to set guide rules, the regulations aren't there? And so, while we don't do that, we ask folks not to share that, we don't share a sense of information that maybe we get from a partner. We don't share personal, sensitive information, these tools when we're prompting. And then finally, and this one gets a little challenging focus on automating tasks. You understand? Well. The New School: your ability to use these tools actually comes from a place of being able to converse with them, to direct them to focus them and actually over time, inform them about what you want to get out of it. And so it's better to focus on tasks that take you a lot of time, but that you understand. Well, instead of trying to do something you've never done before with these tools. So with those guidelines in mind. I'm going to share a little bit about how I was using a tool from Google called Bard to solve an urban climate action problem for myself earlier

The New School: in our work around analyzing interventions to mitigate extreme heat risk, we care a lot about what cities are actually doing. We care about. Where are they planting trees? We care about where policy changes often, and especially in places where I don't live, that data can be hard to access. You can talk to policymakers. They themselves sometimes struggle to track these things. Work is happening across so many different departments, so many different places, and so simply to get together a list of sites where certain actions were taken by government that may add up to better climate outcomes and lowering urban heat can be a challenge. The New School: So in this sort of demo problem, I decided to look at a series of press releases from our own Nyc Park system. These are press releases they put out when they do a renovation. When they build a new park. There are hundreds of them online if you go there, and each of them is written as a press release. Right? This is so and so, said this so and so cut a ribbon, but they also have a lot of very interesting information. They often describe how much money the project costs.

They describe what the project was where it was. And there's a lot of information baked into these press releases. They're also real time. If something happened today, if someone's cutting a ribbon on a park, there's going to be a press release about it. So it's very current. The challenge is extracting this information and getting it into a format where we can analyze getting it into something where we can put it on a map, or where I can actually work with our teams and understand, say, looking at satellite imagery, was there any meaningful change after a project happened. The New School: And so the way this works is, I take those press releases plus the bard tool. The New School: feed it in and give this a prompt I'm sorry. This is text a little small for folks in the back. But this all starts with asking a question. The New School: and I set this up for Bard by saying, I'm going to give you some press releases, and I want you to pull out 4 pieces of information from each of those press releases. What's the location of the park.

The New School: What work was done at the Park? Was this a renovation? Oh, was this entirely new park, etc. And I say, I'd like to know the cost of the work. And then, finally, I'd like to know if this work is focused on improving resilience or adaptation. I want you to note that for me, and then I also ask, is there any questions? These are conversational engines, and they work better when you converse with them. We found. The New School: Bard responds that works. I understand your request. I think you can do that. Let's try this out. The New School: So I feed in the first, and it's abbreviated here. But essentially it's a big block of text that I just copy pasted from a press release and put right in here, and Bard does a very good job. Well, he appears to do a very good job right on the service here. We now have a table, says Shortage, Park and Bay Ridge. The work that was done is a new dog run cost a million dollars.

The New School: And yes, there's been some work here to adapt the surface for excessive rainwater, and it goes into a little detail. They're actually a very useful blurb. So that's great. But to the point of always checking our work and making sure that we're going back and validating what these Llms are doing. Let's actually go to the press fleets itself. So it got the location right. Got the work done right. The New School: It got the cost right, and it even got the resilience. Right? Resilience part right? And it's useful, you'll note, and this is probably very small at the bottom here. But the word resilience is never used, and that's the power of these Lms. They can actually parse texts in ways that are human. So if a human analyst were to read this The New School: and say, yeah, I would score that as resilience. There's an opportunity for the Lms. To actually mirror that not perfectly and not always, but to allow us to pull out concepts and things and categorize them appropriately. And so, of course, one press release is useful. That saved me a little bit of time from reading that whole thing and coding it, putting that data spreadsheet. But where the value comes in is being able to do tens of these or hundreds of these. And so following that pattern again, just looping over press releases. The New School: I was able to build a table, a spreadsheet of this over time. I actually realized that I wanted Bard to actually change something. So I wanted to give me a binary variable of just yes or no. Whether there was a resilience project, and I also wanted the text. I wanted a short summary that was as simple as just saying, Hey, can you go back? Reanalyze these things and add in just a column that says, Yes, no. The New School: Now for any data scientist in the room. And I'm a data scientist

The New School: to do all this normally is a lot of time. I'm sitting there coding. I'm looking up. How do I do this? Especially in my day job? Now, where my main job is meetings and working with people. So if I get an hour to code, I'm sitting there, you know, instead, simply by working the prompts and working in conversational language. I'm able to get this complex data structure together. And of course, back to the original goal. In this I find ones where there are projects that are intended to mitigate urban heat island effects. And so in practice, I would take those, build a separate data set of those, mark those, and then actually look at over time how temperatures change in those locations. And then we can actually look at at scale. Is this working? Can we adjust strategy. Do we need to engage policymakers on a whole new strategy over time? The New School: Interestingly enough, I want to remind you, Lons aren't perfect. These aren't tools that are that are magic or that are doing anything beyond really complex matrix maths across massive data sets. And so it's important to keep humans in loop. But they are useful in being conversational. And a trick that I've used often The New School: is, you can actually ask an Llm. What it's not confident about what maybe you should check by hand. And so in this case there were a couple where, whether a resilience or an annotation project is a judgment call even in practice. If I had an analyst on my teamwork on this we would have conversations around. Is this project an appropriate resilience, project or not, etc. In the same way I can say, Hey, are there any of these that you're not confident about where you read something in the description? But you weren't sure if the goal was resilience or adaptation. And so it actually picked one out The New School: and said, I'm not super sure about this. It says there'll be a new synthetic turf field with a stormwater management system. But I'm not certain if that directly connects to a resilience glor outcom. So in that case I could take that one. I could call up a contact, or simply go to the place or consult with an expert to figure that out, but allowing the imperfection to stand in our way of using these tools actually means we won't actually find out about them right? And so within, with these guide guards with safe lines, we're able to get actually a lot out of these tools. The New School: So in the end. And this is where, as a data scientist, this is super exciting for me. Because now, instead of writing code to geocode something when I had this table. Didn't even have addresses or locations right? I had a park, and then I had A borough or a neighborhood something like that. And I, said, Bard, can you geocode this and then put it on a map.

The New School: and there in 10 min. I've taken text data. The New School: I've done probably years of Nlp to kind of train up on getting Nlp to extract entities to summarize key terms. I put that in a table. And then I made a map simply through about 5 or 6 prompts in a little iteration. And so for us. The New School: at the level of insight and action. This is where Llms are actually incredibly useful. We're getting to insight quicker. I can actually understand something. So I can move on to my big question, which is, how are these projects and implementing them in Nyc parks? How is that actually going to help with mitigating urban Heat Island effects?

The New School: And then action. Well, this, of course, doesn't imply any action directly. The fact that I have a map this afternoon today means that I could hop on a call with a partner or a policymaker much more quickly, and give them guidance and overall. When we think about these timelines of 2,030, we think about these big things? The New School: It's all time. How can we compress that? How can we tighten this up a bit so that the real work, the stuff that will take a lot of work and can't be automated work with communities work at the policy level that that can actually have enough time to do that work because we're automating some of this back office stuff. So thank you for going along on this journey with me. We're really excited to continue using both the Aes today and then the AI in the future. And we always love to work with new partners and collaborators. So please do get in touch. Thank you. The New School: Get you set up.

The New School: Thanks. All right. Let's give both of our presenters a round of applause from Wri in the new school urban systems lab. So I'm so excited that now we're actually going to transition to the panel discussion. And we've got a panel that includes an incredibly impressive group of leaders from across sectors. The New School: and I'm going to briefly introduce them. But I'll let their their remarks do all the talking. So first we're going to have Rit argwal, who is the New York City's Chief climate Officer and Commissioner of Environmental Protection. We have Gino Van Vegan, the Secretary General of Ickley, local governments for sustainability. We have Jaya Dinda from the Wri. She's the interim director of the Wri India Ross Center for sustainable cities. And then we have Alex Diaz from Google Org. He's the senior manager of AI for social good and crisis response. The New School: The panels can be moderated, as Dr. White said earlier, will be moderated by Joel Towers, who's a professor of Architecture and sustainable design at the Parsons School of Design and the Co. Director of the Tishman Environment and Design Center. So without further ado, please join me in welcoming these leaders to the stage. and, as I mentioned before, we are going to have a Q&A session, so please get your questions ready, and I believe there'll be mics that will be brought to the audience to facilitate that. So, thank you.

The New School: Thank you, Josh, and The New School: just to pick up where where Evan left off. It's all about time The New School: and and and the compression of time, and we're running a little bit late. The New School: So I'm going to do my best to compress the time so that we can actually get to a couple questions if we have them. But we have this extraordinary panel here, which I'm very pleased to be able to moderate and following time, and and Evan's presentations. The New School: We have a lot, I think, to think about and to discuss, as it relates to the expertise of our guests today, I wanted to just frame the conversation very briefly, as each of you consider the questions that we'll discuss today around issues that have come up already in the first part of the session around Hope and Fear. The New School: because I think AI generates a tremendous amount of hope, and it generates a tremendous amount of fear about questions of trust The New School: where information comes from. I personally trust every press release that comes out of the city of New York.

The New School: And so, if the underlying data is coming from some place, how are we ground truthing The New School: the information that produces all of these new maps for us to think about? And how does the technology itself The New School: help us advance questions at this pace, and with the urgency necessary to the changes that we know need to occur in the context of a tremendous amount of misinformation. The New School: And so we we sit, I think, at a really important hinge point The New School: in the ways in which technology can help to achieve the the critical goals, the existential questions that time mentioned, and at the same time having to do that in the, in an environment in which the building of trust The New School: seems so critical. So I think, Red, if I could start with you in that context. The New School: because I think New York has been presented here rightly, as doing a tremendous amount of work over a long period of time to try to address these these crises, but knowing you and the work that's going on, there's so much more to do. The New School: you recently did a piece in the Times a couple weeks ago, reminding people that panic and despair are not necessarily useful The New School: in relationship to thinking about emergencies. Tell us how to avoid panic and despair. What are the issues in New York where you see AI and new technologies really helping to consider this future?

The New School: Eric Schmidt? Well, thank you, Joel, and thanks to the new school and all of the organizers, and Timon and Evan, and of course, particularly Timon's work at the Npcc. Which is so important in New York City. The New School: I guess, to to answer your question. I mean what we've seen today, and and the tremendous work that's gone on, and I'll speak selfishly for New York, leveraging data to understand the challenges we face and and help prioritize. The work has been tremendously useful. And so that is a a source of hope. The New School: As I sit here, though I cannot help but think about the flip side of that which is the fear The New School: and to me, and I think the real question. And I'd love to

The New School: put the challenge out to so many of you who are engaged in AI and and data. And the climate crisis The New School: is to think about the delivery because The New School: one of t

2023-10-15 01:49

Other news