 So you have heard exciting ideas and the project plan we put together for this topic AI for clean energy and climate resilience. You heard the students talk but when we execute this plan we realize we cannot do alone. We have to collaborate and the collaboration need to happen at different dimension. Public-private collaboration, utility sector or energy sector with IT sector collaboration, domestic collaboration and also international collaboration. So with that we have a very interesting panel and in the first afternoon session talk about innovation through collaboration. I'd like to welcome Pamela Anson from Department of Energy to the stage. Thomas Bayer from Yang and Guha from Google. Okay, the format for this session each of them will have about five to ten minutes to share their perspective on this topic. They may highlight a very interesting initiative on their organization. They may share very challenging problems they have on their organization. Then after that we're going to have we prepare a couple questions for them but I try not use my question. I look forward to have a question from you guys as well. As I mentioned earlier this is your meeting and we will start with Dr. Anson. Dr. Pamela Anson is a director of artificial intelligence technology office at the Department of Energy. Pamela? Yes hi everybody and I want to say thanks for having me here today to speak with you a little bit and talk to you about what's going on at the department. I am the director of the artificial intelligence and technology office. It's quite an honor to be in that role and I just want to share a few things with you about what's going on. But I tell you so far I feel like I'm in a setting where I'm around peers and people that can relate to some of the things that I'm most concerned with so it's been quite refreshing to be here. I am very focused on responsible and trustworthy AI and I don't see that as a expression or a cliche. We take it very seriously. When we are looking at using artificial intelligence for making sure that energy is distributed fairly, that it is distributed equitably, that's what we mean when we say responsible and trustworthy. So it's a bit more than something that someone is just saying or a piece of a principle that's put in some legislation. So the secretary of energy she has a perspective on artificial intelligence and I just wanted to point that out here. I'm not going to read it but it is clear that she is saying that it is necessary to save lives and it's a matter of economic and national security and we all know that and that is why this is so important and that is why AI and the use of AI for mitigating the risk associated with climate change must be adopted. It must be used in a responsible and trustworthy manner. So I am not going to read these but I will just say that at the forefront of everything that we do is first of all the executive order on climate change and the requirements that are there and the sense of urgency. But in addition, when I'm dealing with and when I'm talking about responsible and trustworthy and the use and adoption and the thoughtfulness that needs to go into building AI solutions, there are three executive orders that come to mind. So the promoting the use of trustworthy AI, the advancing racial equity. So that was the one when I first started talking to people about AI and responsible and trustworthiness and then I tied in this executive order and I would get these blank stairs like what and that but you've heard this all day. You've heard it today, you heard it yesterday and then improving the nation's cyber security. So those are just three executive orders but it overlays with everything, right? AI is infused in everything. So real quick, the vision of my office, I direct the AITO office is to transform the department into a world leading enterprise when it comes to the research, the development, the discovery, the delivery, adoption of AI. And so we're here as coordinators and advocates for the office. We're the glue in the office that helps pull the organizations together. We believe that the sum is much much greater than the individual parts and that's a tough job but that's the responsibility of my office. And so we're not saying AI is the answer to every energy problem but we are saying that where it makes sense to use AI that we should do so and in a responsible way. So this is an image of my office. There I am sitting in the center there and there you are. There is the program offices within DOE and we are the convening organization and the drivers for, we are the catalysts for change and we are there to pull the organization together, look for gaps, look for issues, look for open space and then help us bring solutions to bear in an organized fashion. And then I won't go much into this because I want to save things for the panel itself but these are some of the things that we're focused on. Strategic planning, so I'm responsible for the DOE's AI strategic plan and that's under works. We just stood up the AI advancement council. We're very focused on workforce development and making sure that the skills are where it needs to be and that's one of the concerns that we have and then the responsible and trustworthy AI. So these are the goals and we're on track with hitting our goals and objectives and the important thing is to tie it to energy and climate and be able to help folks understand that we're not implementing AI for AI's sake, that this is all about impacts and driving results and solutions that matter. So we're real focused on going from research and development to actually actual mission outcomes. So I won't go into the council here but this gives you a sense that it is five council members. They are the highest, some of the highest officials within the department, the Undersecretary of Science and Innovation is a co-chair and the Undersecretary for Nuclear Security is a co-chair. And then I sit on the council and then the general council is there as well as the intelligence and counterintelligence. But to counter or to compliment the council is the program committee and that that's the program offices and representatives from the program offices in the department. And the last thing I'll point out is again very focused on outcomes. We are we have stood up the responsible and trustworthy AI task force. We are very focused on establishing ethical principles and equity principles and practices more so than principles, but we need both. And then we're looking into next generation AI. And we can talk about that during the discussion if you're interested. And as I said, I'm working on the AI strategy. The team is on what we're done is going through the vetting process. We believe in innovative governance. We believe in advancing the AI ecosystem. And I won't touch on this one. I'll leave the slides and you can take it. But this is an example of one of our outcomes. In addition to the council itself, another example outcome is an integrated development environment. And honestly, this needs to happen in collaboration with industry, academia, independent researchers, searchers, as well as other government agencies. So not just an integrated development environment, but also equity hubs. So we've been there's a there's a real push for establishing equity hubs so that information is represented. And so that led into the next slide, which is speaking to distribution of solar is one of the concerns of one of the value propositions of equity, and why equity in the energy transformation is so important. We don't want to exclude communities in data set. And that's easy. It seems easy to do. So we find that to be a problem. deep fakes, content authenticity, got to deal with content authenticity. If we are going to make a change when it comes to accelerating the risks associated with climate change, we need content authenticity and then balancing the science and reasoning. That's where AI comes in as well. There must be a balance. So it's a social type of technology. And that's basically what I wanted to go over. These are just examples of the energy grid, the predictive maintenance, home energy management, all examples of applied artificial intelligence to make a difference in the day and time of the energy transition and the transformation. And the last thing is on my mind is what I just said. I'm very concerned about the fact that AI is underutilized when it comes to climate and grid resiliency. There is such an opportunity to do more. AI for energy transformation is necessary. I think there needs to be some type of consortium to that extent. And then we need to we need more significant data management under way. Okay. Thank you, Pamela. I know some of you may have a question. Let's hold the question a little bit until we finish another two quick overviews. So many of you, some of you were here on Monday and have heard the five-side conversation between County Rise and Ruma Jinda regarding what's going on Europe and the importance of energy security issue. Now we are very glad to have Thomas Bear, is a senior VP and the chief innovation and the strategy officer from IAW, which really give us the first-hand perspective and opinion. What's happening in Germany? Yeah, thank you, Liang. At least I give you first-hand information of what is happening in our company. And yeah, I would like to share a quick snapshot about the IAW strategy and appropriate to the great auditory here today. I would like to highlight in particular our chances and challenges which come along with innovation, technology and partnerships to make the IAW strategy happen. And on that note, we as IAW, we are founding members of the Bits and Watts program and I want to thank you, Liang, and the entire Precord Institute for Energy for the great collaboration and the inspiring impulses they are setting. It's really amazing. So before I jump now into the matter, let me share one pre-remark on a personal note, a quick statement about one concern which is in everyone's mind seat stays. We are all struck by the terrible war in our neighborhood in Europe, the war in Ukraine. And our compassion, of course, is with all the victims who lost their lives, their beloved, their homes. And we as EON, especially our colleagues in Central Eastern Europe, which are all really almost at the front line with their business in Slovakia, in Poland, in Hungary. We are trying to do whatever we could do to help, right? And it's almost a second order problem when we talk about business impact there. But I got asked quite frequently those days, what is the point of having a strategy at all in that moment? Aren't we back on square one with all the energy transformation and stuff like that in light of these terrible events in Europe? My answer is, and that is what I want to bring across today with my short talk, it is even more valid and more necessary and more powerful to execute a transformation strategy now. What we see today in the market, policy makers, regulators, that is all a resounding confirmation that our way towards energy transformation is really, really is right thing. But let's get into the matter. Just a few words about our business portfolio, because that's also important in that context. Our strong pillars in energy networks and customer solutions make us one of the largest European energy service companies, straight in the middle of where energy transformation happens. Our distribution network, which is the largest in Europe, with some 1.6 kilometers of lines, is a backbone of the transformation, because energy transformation happens in the distribution grid. We connect 800,000 renewable plants to our networks, and this is a number which exponentially grows literally as we speak. That is a journey which comes along with a strong move towards digitization, a necessary prerequisite for the whole transformation to happen at all. And I come back to that in a minute. Our customer solutions support our 50 million customers in 15 countries in Europe to become part of the transition themselves. We don't own large generation assets anymore, hence we are a pure downstream in energy, and we are driven by a strong vision to bring good energy to the people and to make energy transformation happen. And that is based on a set of core beliefs, technological and societal core beliefs, as of what will happen and what needs to be done for getting to net zero in Europe until 2050 at the latest. We believe in electrification of industry and society. We need to expand the renewable production on the continent as much as we can, but that will still leave us with almost the same amount of energy which still needs to come as imports to Europe if you want so as green electrons stored in molecules. And we will see these energy carriers converging all over the place. And then we will face an energy system of the future which will become simply too complex to be managed by only putting more copper into the ground. The word of tomorrow will simply not be able to control it by human brains only. We need to radically change the way we do our business by consequent digitization and employment of emerging digital technology to the utmost extent to get to net zero while keeping the lights on. So, I mean, if you forget everything I'm talking about today, which I hope is not the case, there are three things which could stick as a core paradigms of our strategy. First of all, it's sustainability as a value generator for business and not as a bureaucratic burden. Digitalization as a true game changer of our business and growth, sustainable growth driven by the sector transformation and our role therein in the first place. So, these three things, this is a very busy chart which explains a bit why sustainability is such a huge driver for us. The bottom line is, first of all, our infrastructure, our grid infrastructure enables all other market players in Europe to get green. Putting a windmill somewhere which is outpouring electrons is not a big deal anymore. Bringing the right electron at the right time to the right place certainly is. We have more than 30,000 industrial sites under contract where lowering the CO2 footprint is a technological challenge everyone is dealing with and we deliver the products for. And we build the infrastructure and customer applications of tomorrow on our journey towards green gas and the hydrogen economy. And just a snapshot, the billion dollar question is how to get the heat green? That is a big unsolved. How to decarbonize the hard-to-abate industries? Those who apply high temperature processes like chemistry, glass makers, ceramics and what have you. And how to even convert the households in the big cities getting green, right? And coming to digitization also a very busy chart I could talk about for hours but the bottom line is decarbonization means electrification of society. Electrification means end-to-end digitization throughout the whole value chain. And I skip the details of the digitization program and I come to some examples. A quick one here is how digitization enables us to transform a very hands-on physical process into digital. It is the vegetation management. I learned a bit that this is an area of interest here in California. You know we have to cut the trees and bushes around our power landlines just for safety and maintenance reasons. Which in the past we did rather harshly, right? Chopping all green down to the roots just to be on the safe side. Now the target is to cut as much as necessary but as little as possible, right? To do so we use digitally analyzed LiDAR data to determine the right amount of cuts to be made to manage our subcontractors. Their work will be verified by satellite imaging and computer vision techniques and they will be used to identify critical vegetation which is coming back. We already do that in Sweden with 8,000 kilometers of landline and we're going to roll it out to out the 50,000 kilometers of landline of our group. That will save 700,000 tons of CO2 at the end and that's not a piece of cake that really moves the needle, right? Another example is the Ion meter worse. We will launch it June this year which is the first of its kind in the energy world. That will be a completely new digital experience for customers, employees and scientists which will fit with first value generating use cases early on. Imagine the live demonstration for example of an industrial customer to walk him through the alternatives of modernizing his CCGT or heat recovery system or waste circulating or something like that or we set up a carbon exchange offset system in the meter worse. Many ways to visualize complex data in this space and the opportunities of this new experience space are almost endless, all right? So last but not least what means growth to us. Just in a nutshell Ion will invest 27 billion in the next five years into sustainable growth mainly fueled by the energy transformation. Exponential growth of renewables in all our regionalities we are active, new data warehouses, gigafactories growing like mushrooms they all need to be connected that is I mean don't quote me for that but that is growth you cannot you can hardly run away from that is really that is that is something we have to do yeah. We distinguish between areas in our core with steady growth over the years these are the growth engines but in some areas we call exponential wildcards the bit out of the EV charging infrastructure in Europe a new digital platform for third-party access to grid services behind the meter and before the meter, B2C solutions for residential customers who also want to progressively be part of the green journey. All right let me let me close here the point I want to bring across there is a completely new game of energy emerging and we have the impression that will be our game as downstream utilities as utilities with a vision with a vision of bringing good energy to the people and this transformation ahead of us is already real and happening and it will offer huge chances for us. The key to success is innovation, technology, entrepreneurial spirit and partnerships on that way partnership with other companies from startups to industrials we have participations in more than 50 startups all over the world but also with scientists and with academia with our academia partnerships so thank you very much for it for your attention. Thank you Thomas so we have heard the watts perspective then I think with bits and watts you know naturally we'd love to hear the bits perspective and Jay Prickle asked on Monday you know whereas IT companies we should engage IT company on this energy transition so now the answer is here so we are very glad to have a Guha here fellow and the VP of Google but another thing he is doing right now is the founder and the leader developer for the data comments Guha. Thank you Liam and it's really great to be here about five years ago I was standing here talking about this thing I was going to head to you know want to get built and Arun pulled me aside and said jump in the trenches and start building it yourself and so five years later here's the report so to set true is to set the context there's a ton of data out there from all of these different entities this data is essential for everything from science and journalism and policy and using this data is unfortunately incredibly painful involves data wrangling which is repeated over and over again there's an analogy with satellite imagery in 1998-1999 NASA had its landsite imagery upon its website it was so difficult to use almost nobody did Google Earth and Maps came along did all this data wrangling once and for all and literally changed the way we look at the world around us today if you want to ask a question like which California counties are more at risk from climate change well the data to answer that exists but it's so difficult to use in fact the first step in answering the question is to write a grand proposal imagine instead if you could just walk up and ask a natural language and start beginning your exploration so what we've been doing for the last almost five years now is we've taken a very very large number of data sets we've done all those data wrangling we've cleaned it normalized it aligned the references build schemas on top of schema.org and build this giant knowledge graph it's too big for most people to download so we provide APIs on top of which different kinds of applications can be built the four target audiences consumers so can just ask a natural language policies and folks and journalists what I call non-programming researchers dashboards and visualization tools for them application builders and one very special community of builders which is people building AI models to do different kinds of things there's three pillars to AI algorithms compute power and data if there's no data there is no AI and that's why data commons comes in in this context so what we've built one is an open source infrastructure for creating and storing and building these knowledge draft visualizations there's integration into Google search engine but 90 plus percent of the work is elbow grease in terms of demographics economics health climate energy food crime it to give you an idea of the scale it's about three billion time series it's about four times the size of Fred which is a federal reserves economic database that they use to run the country's economic decisions you know just some examples of that you can go to Google Google today and type in these kinds of queries and get answers everything from energy use for capital and India do incredibly complicated questions like number of poor Hispanic women in Santa Clara County a driving application for the last two plus years has been sustainability climate change is happening and it's not as simple as 1.5 degrees versus 2 degrees versus 2.5 degrees climate delta is very widely this is in all the visualizations I'll be showing you today is all based on data and visualizations from data commons.org this is temperatures as predicted by one of NASA's popular models in 2050 related to 2006 according to RCP 2.6 which is the most optimistic scenario and you can see that even in this ultra-optimistic scenario there's going to be places in this country where the temperature goes up by more than seven degrees centigrade there are other places where it goes down by four and a half degrees centigrade to use Aaron's analogy it doesn't matter if you have your head in the oven and foot in the freezer it's both you're in trouble and that's what's going to happen the other point is that there's many existing inequities everything from hunger and poverty and diabetes and health insurance and the list goes on and climate change will worsen these inequities and we have to prepare and in order to prepare we have to know who's going to be most affected 10, 20, 30 years from now and in order to figure out who's going to be most affected we need data not just about the climate but data about food, health, farming, water, employment and so much more these are very messy data ecosystems just at the U.S. federal level the data about this data is distributed across so many different agencies data commons mission is to organize this data and make it universally accessible which you might notice is Google's mission and because this is so important Google is doing this in an open fashion not only is the data open the software stack is open the entire process is open and so that everybody can participate everything we do is on Github you can download our code you can download the data and we do expect there'll be you know companies and others who build applications for profit on top of this but this is so much elbow grace it's kind of pointless to do it over and over again so we should just do it once and make it available everywhere many many many topics have already been covered everything from all of these topics and many many more and they're coming in at an ever-increasing rate but they give you an idea of the kind of stuff they can do let me just walk you through a few examples we all know that with heat cardiac conditions become worse this is one of our visualizations each dot here is a U.S. county and the x-axis is the x-axis is the expected temperature rise the y-axis is prevalence of cardiac conditions what's interesting are the places the counties to the top and right if you look right at the top there there are counties in South Dakota there are counties in New Mexico and so on which have a high incidence of cardiac this thing which are going to become much worse and unfortunately if you drill down you'll realize that so many of these counties are the very Native American reservations are in many ways what we need during the pandemic we relied heavily on the dashboards from Johns Hopkins to figure out what was going on for the next 10-20 years we need dashboards like that except a thousand times more sophisticated thousand times more complicated not just dashboards platforms that we can run our code our models our analysis on and this is exactly what we're building this is you know irrigated land versus projector temperatures to tell you what kind of places which can have most kind of effect this is and of course we collaborate a lot and our first and most favorite collaborator especially now that we hear is Arun and this is IPCC temperature this is the wet bulb temperatures and it's kind of one of these places times where I wish we were wrong and India actually hit those temperatures last week not good there's one last important point which is the model is not one data commons one ring to bind them all it's much more like websites we have many different websites and all these different websites share as common schema which is html and a common API which is Http some websites are behind firewalls some behind paywalls some are open but they're all you don't switch your browser when you go from your intranet to the public web and your intranet can link to the public web the same model over here which is that you can have these private or other data commons and the analog of linking is joining a great example of this is you know IIT Madras the best IIT of course you know set up a data commons which is India focused and they're especially focused on water issues so the the black dots there are the glaciers that are feeding the Indus, Ganges and Brahmaputra which together feed a billion people and when the temperatures at these place at these points go up by more than a certain extent they stop being able to hold water and then we're going to have a billion people potentially start and preparing for that is going to take on the audit of decades coming back feeding America is another one of these partnerships feeding America is an NGO in this country which helps run about 2000 food pantries and they have developed this interesting index for food insecurity and which takes us to you know you can go to data commons.feedingamerica.org and to go back to the first question I asked which California counties are more strict at risk from feeding from rising temperatures the x-axis is their food insecurity index the y-axis is the temperature rise and you can see it's the interior agricultural counties which are most at risk to end on a positive note there's a solar energy potential versus poverty there are places which are very poor but which have high solar potential where we should be investing more as I said we like to love to collaborate it's all open go to data commons.org collaborate more with us thank you thank you Guha so we have heard that California PG&E perspective this morning you have Australia have Germany and Guha mentioned a little bit on what's going on in India so my first question to all of you is really about all the plans to be net zero by certain time whatever it's 2050 or 2060 or even longer and because last weekend, Saturday many of you may realize or may not realize the whole California grid was running on clean electricity for two minutes in the late afternoon April 30th okay and so do you think in Europe, in the United States or any other countries can we reach net zero by 2050 or by 2060 or longer or to start from this end from Thomas to Pamela then to Guha first give me a quick answer yes or no certainly yes okay Pamela I think yes by 2050 good good Guha I don't think I'm qualified to answer that I think it's much more complicated questions than that okay then that's the next opportunity you guys have if we can reach net zero by 2050 what should we do what's the risk we are facing what's the transition energy transition risk we are facing especially under the situation like in your region the war is going on and also in the United States has a different change of administration and different states the state federal coordination and also the what's the role of IT company let's start from again from Thomas well I mean I guess the biggest challenges are not so much on the technological side although they are huge but they are in my point of view all manageable my point of view the biggest risk on that way is the challenge of the society transformation we are facing yeah you at the end of the day our continent is or the European community for example is 27 countries which means 28 regulations very different mindsets on what is the role of politicians and regulations and what is the what is the role of the industry and we see a lot of let's say counter balancing action right now in the crisis what we see on the positive end is a rebalancing of what we call the energy trilemma the good balance between sustainability affordability and security of supply that gains a lot of importance what it did not have in the past years the flip side and the dark side of it is a lot of politicians are overshooting with over regulating the markets and that in my point of view could be a big setback on this way you know clearly I definitely think that I definitely did you hear me did you hear something he said we can drill the mic it's good yeah so I definitely think that to to get to where we want to be by 2050 I know our goals are more aggressive than that in the United States but for 2050 I think it's possible and the reason why I say that is because we are paying attention to the climate change we are paying attention to infrastructure and grid resiliency and so as long as we stay focused on that and also remember our neighboring countries because again there's a ripple effect so as long as we pay attention to that and our action oriented which someone talked about being actionable earlier today and if you know me at all I'm very much about stop talking let's do and I believe that from a climate crisis perspective that we are doing something about it I do want to say something real quick about artificial intelligence if it's okay so AI can either I was in this discussion earlier with some students and they said just use the word amplify so AI can amplify good and it can amplify not so good as long as we are paying attention like we are now we're looking at how we can use AI to better solutions to address climate change to address grid resiliency to look at how we can accelerate and process the magnitudes of data and the different data types and as I said earlier I don't think that's the sole answer but I do think that it's an instrumental part of mitigating the risk of climate change we have to know and I think we're we can do more of this but I do believe by 2050 we're going to be there we have to be able to anticipate crisis before they occur and they don't always have to be crisis that have already occurred because that's past data and that's old so we're going to have to be able to anticipate that that we don't know and that is what the that is what happened with that is what's happening in some of the communities and some of the crisis that we've had we didn't know how to deal with it because we weren't prepared so we're going to have to prepare ourselves and AI can help us with that so the artificial intelligence the simulated data as well as real data based on historical experiences can help us with that and the last thing I'll say is your links so if there's a chain there's links in the chain and all of those links is what makes the chain a chain every community matters equity matters equity until we pay attention to equity and ethical responsibility we're not going to get where we need to be when it comes to the climate crisis but I think we're working at it and I think it's possible Guha you want to add your perspective or you want another question from me yeah different question very good so that's a question for you to continue what Avila said okay I think if you guys would love the the AI question more then yeah sure let me answer that question which I think she answered which is what are some of the near term things that AI can do and I'm going to go incredibly specific there's a lot of problems which are have essentially computationally unsolvable everything from you know grid optimization OPF and to you know actually solving the navier stokes and all of these kinds of things I'm really excited by more recent solutions which don't try to actually solve the problem but use various deep neural network techniques to get somewhere in the neighborhood of a solution and you go from you know like I forget the name of the student who showed four orders of magnitude improvement right that actually gives us the opportunity to do that together with sensing technology and so on and so forth and change the name of the game in many many many different dimensions that's absolutely huge and we're going to need to bring that in to reduce so we look at the in order to go net zero we have to be able to reduce the cost of producing a calorie of food right and that's not going to happen the way we are doing it right now if anything the cost of a calorie energy cost of a calorie is going up as the consumption of beef is going up how do we address those kinds of issues in material science there's long-standing problems which you know many people including people in errant's lab are making some progress essentially doing a a variant of using deep neural networks to investigating incredibly large search spaces so it's positive and I don't know enough about those material science or food stuff but the one place we can help is in many many many of these projects grad students spend 70-80% of their time bringing together the data set cleaning it up doing it over and over and over again if they could instead operate on a common platform so that one of my favorite examples on this is what's the relationship between poverty and in unemployment and hypertension and obesity well that would be on the order of a term project kind of size project given that the data comes from four three or four different federal organizations and so on once you can assume you can open a python notebook anywhere and assume the data just exists with a common API it's now being done in Berkeley in under an hour by 400 plus students for a grade and if you can bring that level of advancement and that level of speed essentially if you can empower every student out there to do things four times sorry four orders of magnitude faster interesting things are going to happen okay I before I let you go follow on question let's ask from your side to Thomas is really what's your view of the level adoption in the power and energy sector I pretty much know the answer but I would like to hear you again you know what's your view because you've been dealing with a many industry you're dealing with health care dealing with a you know social network consumer behavior exactly what's your view of Guha the level adoption of AI in the power and energy sector let's start from you then with Pamela and Thomas let's use this way elementary school high school or university I'm not qualified enough to answer that but based on my conversations with Ram Rajgopal and Arun there is let's say room for improvement same room for improvement good just just a a quick quick anecdote when I started in the utility business in the very early 2000s my colleagues told me you know more than four percent of renewable fluctuating renewable energy in the grid simply won't work it will bring it to the collapse because the grid is we have such a central focused architecture it will let's say create some swings and the whole system collapsed I mean we have hours and days in in in in northwestern Germany where let's say over hours the whole system is supplied by renewable energy only in the meanwhile so 20 years later we have in average in Germany 50 percent of renewables in the system and we are managing the grid literally by and large with the same methods as we did 20 years back there by now there is not very much emerging digital technologies in it and the question we all ask ourselves when do we hit the iceberg where is the infliction point where it simply doesn't work anymore and we have we have the impression we are very close to it that the system is simply not manageable because of the sheer amount of electrons in the system and the sheer necessity of creating self-organizing systems where the point of generation of electrons and the point of consumption needs to come closer to each other really demands these technologies so a clear answer we will see AI but also distributed ledger technologies augmented reality all these technologies throughout the whole way you change I'm 100 percent convinced terrific let me make a pause here and see question from the audience so Gula as you were describing the data comments I was just smiling and thinking ooh new toy sort of sort of thing but with such incredible information indeed to guide decision making and address new questions in particular related to sustainability issues my question to you is the data comes with different levels of quality and so how is the treatment of uncertainty across multiple data streams reasonably addressed so that that would be one and the second one comes related to Dr. Eason talk the part on responsibility a lot of the things in terms of visualizations are correlations and not causations so how to address those two layers importantly as this effort continues to grow yeah let me answer what questions are very great questions thank you the first one is you have data from you know everything from the U.S. census to some state in India and there's often so what the short answer is a we maintain the provenance of every single data point all the way down to the code that was used to clean the original data source all the way down to the original files that it came from so that you have transparency and then second level we have four surfaces like Google search we use only a small number of the datasets that are sources which are sort of hopefully beyond approach like the world by U.S. census things from peer review journals and so on and then finally for researchers you can actually filter your query by specifying you want only data from these sources the second question you raise is a much much much bigger one and it's a debate that we've gone through which is yeah you can look at a certain in a scatterplot and say that's a causation it's not it's something else is going on is the solution to say no you can't do the scatterplot or is it to figure out ways of educating whoever is doing it because if somebody who wants to create misinformation will find a way of getting the scatterplot because it's not that we are it's there it's just and so a big part of what we are trying to do is actually work with communication school departments and generalism schools to fund competitions to help people tell the story because at the end of the day charts like that are useless they need to be told enough to a much larger audience of people very few of whom are in this room especially politicians and so on in terms of stories that they can relate to and it's part of that also educate sort of critical data thinking if you will in terms of understanding the distinction between causation and correlation and a whole bunch of these things for uncertainty to add on top of what was said I think that uncertainty so there's a there's a certain level of fear when it comes to AI and then there's why is that the case and I think where we have an opportunity to to build up that confidence is to look at so track ways to identify the level of uncertainty and what to do with the response or the results or the predictions of the AI based on that uncertainty and it all depends on situations it's very situational and so there is opportunity though to really start taking a look at and have the models convey back the level of uncertainty when a a prediction is about to be made the uncertainty level is X and the humans in the loop when they talk about the the human centered AI here they talk about the human centered AI and that's so important because that's where that comes into the picture so that the validation that's needed is infused into the process and we aren't 100% there yet but it's something that's getting attention it is a fundamental requirement when it comes to making sure that if I'm going to use AI to help to tell me how where the solar panels are across the various homes and where are the opportunities to improve the solar panels before they break right so as an example AI is going to give me that insight through computer visioning etc etc I need to be able to count on their data that is giving me back and it is based on the data that it gets but it's also based on how we make the verification the validation deal with the uncertainty and that is something that is getting attention it's a very good question it's something that's getting attention it's like how we deal with uncertainty today as humans how do we deal with it right you you have to and the you we need to think about that and the last thing I'll say is the diversity or the the interdisciplinary teams when we are building the algorithms when we are establishing the models and even when we are establishing the data sets there needs to be that interdisciplinary set of teams involved not just what you're familiar with that and then that's when yeah so that will help with some of that but definitely we have to deal with uncertainty so I deal with Emma with Emma quite a bit for solving Navier strokes specifically and I was wondering how all of you think about generalizing out of sample because ultimately you have to go into the future which we have no data for do you want to go first you want me to go first I'm all for it I'm all for moving out I'm I'm mixed so there is the machine learning element and then there is the more generalized is where I think you're going to and where there's not that much human in the loop but I don't think we're ready yet but I do think that this is this is where we are going and I'm in favor of it so yeah great question I don't think the community hasn't answered for it yet it's just early stages of papers look I got an interesting result I got an interesting result like why or how yeah I fully agree with with you both what I want to add to that question is it has a a social and even a regional component in it as I said we are doing business in 15 15 countries from Sweden in the north to Italy in the south from UK to Turkey and the way people approach technology technological progress data security is it differs widely I mean as an example in Sweden we are rolling out a second generation of smart meters already in Germany we have not even started only due to let's say an almost irrational fear of data fraud and I mean we take data security very seriously but it is it is absolutely bizarre right the the the German way of reading meters is still someone goes in the basement takes a card writes down the meter and and sends it to the utility mixes up the gas meter with a power meter big headache and and it differs between big cities between countryside and so on I I guess our answer is we have to find positive showcases for it we have to to to to take a first move here and there even on our own expenses and and say look this area where we try to employ all what is possible today and have a look and feel and and and see how it works we have to we have to create this positive showcases to convince people at the end of the day there's one point however I'm sorry I'm diving back into the weeds over here if you look at the different climate models right this is like IPCC CMAP6 has like 80 models or something like that they vary across the range of predictions is so large it's like six degrees centigrade across and you look at them and you say wait you're just ignoring cloud cover and you're placing such a huge weight on cloud cover because that's all unknown there's an approach that is being used by two of our own students both of whom I saw earlier who disappeared which is basically say look I'm going to ignore Navy Stokes and all that I'm just going to look at historic trends and and together with CO2 rises etc and then you because we have so many weather stations all around the place to see if that works better because intuitively you could say that those models are also incorporating the impact of cloud cover and ocean temperatures and you know butterflies in in Central Park or whatever I'm not saying it's going to work it is an interesting thing alternative approach to and it's at the end of the day it's like things like Navy Stokes and all of these things are sort of analytic solutions that are based on causal models of the universe and then you have machine learning which is basically curve fitting very high number of dimensions and and I don't know which one I like actually I like to like like to hear your thoughts on this topic okay great this is a giant setup for your talk wonderful that's a question she prepared for herself so okay we have about a couple minutes to go let me ask one easy question for all of you what's the role of university like Stanford and they have you heard you heard from Steve Graham and Ian Arun this morning the plan for the new school we will have a kind of data-centric to help to coordinate all the data related to the sustainability and the climate change because that's how we how can we work with you we need some guidance we can start from Thomas from industry then to the public further perspective then to the IT side so there is a very strong symbiotic relationship between academia and industry at least in our industry so we are energy industry it's not a particularly R&T heavy part of the business we are not the people like the chemistry or pharmacy who put 10-15% of their turnover into R&T traditionally a lot of R&T happens at our sub-suppliers or in cooperation with partners with universities and for us let's say cooperation with universities was always a large provider of impulses and a past maker for progress in our industry yeah I cannot agree more I would like to see more collaboration so this is a great form this is a great way to talk about things and explore what others are doing and then together we should start coming up with solutions together with industry with academia with the federal government with the state let's look at how we come together maybe do a energy transformation consortium as I've been mentioning before or something like that I like the data commons so let's come up with some common taxonomies and solve some problems together so that when we're solving those problems they automatically scale because look at the amount of people involved right it's so significant impacts together so that we all weigh in and buy in on the outcomes good cool so I'd first like to object to your phrase there universities like stanford there's only one and stanford is redefined the way so many different things happen now seriously look at there's somewhere around here there's a photograph of of Terman Terman came up with the concept of a stock option that has had such a profound impact on everything every it's so many different things right and we have and stanford is interesting unlike some of these other universities on the east coast stanford the nature of stanford has changed so much over the last even the last I mean I meant here 30 years ago even the last 30 years it has completely different in its character so we have no idea what all challenges are going to come ahead of us in the next 10, 20, 30 years stanford was one of the first to create this new school right I think stanford can define its own kind of get and redefine what the role of a university is in this context so great I love your answer so let me summarize what we learn the action item to move forward for the collaboration stanford needs to take the leadership role yep and we need to create a consortium which is open innovation can play and we need a more research funding thank you all of you