IGF 2021 – Day 0 – Event #111 Living in the darkness or on a desert? Will new technologies prevent an energy crisis?

The following are the outputs of the captioning taken during an IGF virtual intervention. Although it is largely accurate, in some cases it may be incomplete or inaccurate due to inaudible passages or transcription errors. It is posted as an aid, but should not be treated as an authoritative record.



>> So after this wonderful intro I would like to welcome everyone at the discussion panel devoted to new technologies and energy.  Living on the desert or in the dark will new technologies prevent energy disaster?  That's a question that we're hoping to be able to find at least partial answer to. 

      It's a particular day because ‑‑ I don't know perhaps most of you in this room around the globe who can watch us thanks to digital technologies may not know that in Poland today.  We have a harsh energy condition.  There's a Swedish power plant which was ignited; started just to help the Polish grid, which turned out to be not sufficient. 

     So Germany, Ukraine are delivering energy to Poland, which is a bit awkward because until very recently, like days ago, we were a net exporter of electric energy, so the situation within grid ‑‑ even traditionalists, we had it in ‑‑ Poland can change something very fast, and this is going on and something we have to remember as we start a discussion.

     And I would like to welcome all of you in our audience and also I would like to cordially welcome our panelist.  

     Mrs. Wanda Buk is the PEG group vice president for regulation.  The strange COVID times decided that we cannot have panels like we normally do in the room but applaud for each panelist. 

      So Mrs. Wanda Buk was an undersecretary from state from the minister of digitization at the same time the member of the European committee and the standing committee and joint mission in government and local government.  Now she's working for PGE, one of the largest Polish energy companies as vice president for regulation.

     I would also like to cordially invite the professor Krakow for mining and steel, and sorry for those camera movements.  They were pretty abrupt.

     Specializing in dynamics systems and applications in mathematics in technical sciences as of 2019.  Dean of the faculty of applied mathematics at academia at the Honecha University in Science and Technology in the United Kingdom.  I would like to welcome Aidan O'Sullivan from the College of London.  I hope you're with us, Professor.  How can I check it in this distributed ‑‑

>> I'm happy to be here.  Thank you very much.

>> Oh, you're here.  Wonderful. 

      Professor O'Sullivan is associate professor in energy and artificial intelligence at the UCO Institute where he leads the energy systems and artificial intelligence lab at the department's Data Analytics research team.

      Once again, welcome to the panel.

     From Spain I would like to cordially welcome Professor Juan Guirao.  Sorry for the trouble with pronunciation:  Juan Luis García Guirao.

>> Hello, it's a big pleasure to be here.

>> Thank you to have you here professor. 

     Professor Garcia is a professor is a full professor of applied mathematics at the Technical University of Cartagena in Spain.  He became the youngest mathematician full professor in Spain at 33 years of age.  Congratulations.

     And also from CEE region ‑‑ not just from Poland from CEE region ‑‑


>> Not just me; right? 

>> No. 


>> Well, we can moderate ourselves.

>> There's an issue with the digital panels of technologies being interrupted with internet ‑‑

>> Yes, ironic. 

>> Ironic.  Maybe this is because of growing demand for the energy.


>> Okay.

>> You're muted.  You're muted.  IGF5, you're muted. 

>> Roland, you are muted. 


>> What we usually do when we're in one room is a very fast exchange of thoughts. 

      I'd rather ask a question to each panelist, and we'll be listening to you speaking, and I think our technology today is provoking also this type of discussion, but I'm sure we will overcome both COVID and the technical challenges not issues ‑‑ challenges

      So before we start, I would like to ask everyone in the room and participants in . Digital form:  If you would like and be so kind to enter the web page WWW.mentee.com where you can participate not only in listening to our conversation but also in voting. 

     There are three options available in this vote.  A: Societies should continue using energy systems as of are; B:  Societies should abandon or phase out fossil fuels despite the risk of energy supply limitations or c:  Societies should take a way of durational use of fossil fuels maximizing the speed of development technologies that will allow for future broad use of low or no emissions internet.  In 3 ways we will all be very interested what you think which way societies around the world should take.

     And let me start with first question to Mrs. Wanda Buk.  Your company manages over 5 million consumers in Poland, 5.7 customers and digs over 40 million tons of lignite per year.

>> Yes.

>> Let me ask you:  Do you see alternatives?  Are there technologies within our reach as ‑‑ as civilization or maybe within the piece of PGE, could cut emissions by 2030 by over half in 2050 to 0.  They are very bold regulations, great, or at least ambitious goals but is engineering going to deliver?  Are we going to need by the way, more energy in the future or less thanks to energy efficiency?  Are we going to live in the desert or in the dark.  What's your take on this?

>> Okay.  So, first of all, thank you for having me.  I'm glad that I can participate. 

      You started this intro about the mentioning the very high demand for the electricity for the energy right now, and I conclude from that nowadays the whole Europe and the whole world is looking for ‑‑ is looking for energy efficiency and, of course, we ‑‑ we assume that it will be brought by new technologies innovations and also based on the artificial intelligence but still to do that we have to replace our current generation with ‑‑ when it came to PEG we announced our new strategy.  It's green technology.  We're going to achieve net zero ‑‑ we're going to be fully green in ‑‑ let's say till 2050 but until 2050 also we have very significant checkpoints aligned ‑‑ aligned away to this path to total green transition, which is ‑‑ and this first checkpoint we have in 2030.

     In 2030, you mentioned the number of lignite that we're digging.  But by 2030, we would like to provide to our consumers energy from offshore, wind farms, and we're going to 2.5 gigawatts in offshore wind farms and also and also we're going to invest wind farms on shore, and so there are huge challenges ahead of us, and this we need to combine with ‑‑ with this ongoing electrification and growing demand for the electricity because nowadays when we think about ‑‑ nowadays when we think about the electricity we are not ‑‑ we are not counting transportation, production process and heating, but these areas will become ‑‑ as we are moving away by fossil fuels we'll be covered by electricity, so it will ‑‑ it will also ‑‑ it will also result in growing ‑‑ growing demand for electricity.

     And simultaneously to the ongoing electrification, what I already said ‑‑ we are going to transform power generation to meet the challenges of the carbonization and restricting access to electricity is not acceptable.  I think everyone agree with me.  That we cannot live in the desert, how you said it. 

      So, in fact, it's a complex task that cannot be easily accomplished but ‑‑ and simple ‑‑ simple development of refrainables is not the answer because as we're going to ‑‑ because PEG, as I mentioned, will provide gigawatts of energy during impending decade, it's still not enough.

     Of course, it's on the top of the agenda in PEG to deliver new green capacities, and we lead our core project forward but to provide for reliable access to electricity, these new sources are to be integrated and the balance in the electricity system as a whole. 

      And an appropriate backup is still essential, so that we have to ‑‑ we have to plan the commissioning of fossil fuels based as it's reasonably not ideologically and emotions cannot replace common sense, so we should from our perspective focus on the energy conservation, excellent construction of a our new green capacities and development of new technologies, and one day we'll be able to rely on the radables and storage focusing on a holistic approach will not solve the problem so basically that's my answer to that, and I think maybe I will step in later because I don't want to consume more time, I would also like to listen to your perspective about that. 

>> Just let me ask one more question, which is ‑‑ which needs to be directed to energy sector representative as yourself, do you feel ‑‑ I mean, perhaps making energy production more green requires more distributed way of producing, maybe less concentrated huge power blocks.

>> Yeah.

>> And far more ‑‑ far more small producers, do ‑‑ does energy sector feel comfortable with that perspective all in all even off the sea there's no wind so offshore forms will not provide but farm environment even ‑‑ if nothing particular bad happen small rivers will keep on running forever, but they will not produce much energy, so it's it needs to be distributed.  How much ‑‑ I mean, how do you feel you and your colleagues engineers feel comfortable or uncomfortable technology wise you feel looking at those challenges?  Do you feel you're ready for all of this ‑‑ all of those things.  There's a huge R & D to be done in the next decade or two ‑‑

>> Yes, as you said, you know, the specificity of the radables wind does not always blow and sun doesn't always shine so, of course, what I've already said we need a backup.  We need a backup, and we need to think about a backup and also we need to think, and we need to ‑‑ we need to do R & D, and I think it should not be done ‑‑ it should have been done yesterday and mostly ‑‑ I'm speaking mostly about ‑‑ about the power storage technologies, which is necessary if we want to maintain renewable capacity without any significant backup based on fossil fuels, so that's for ‑‑ that's for sure.

     And R & D while there's a mass of the carbonization targets if they were need to be met, yeah, for sure we should ‑‑ we should focus on that and the whole world, in fact, is focusing on that.  We are all aware that there is a need for power storage technologies, and there's a need how to ‑‑ how to ‑‑ how ‑‑ how to provide backup and also there is, you know, know nuclear power is on the cusp ‑‑ there's not without the reason; right?  And also provide a backup for renewables. 

>> And also a matter of scale of nuclear energy; right?  Huge blocs ‑‑

>> Yes, yes, sorry, sorry.

>> And perhaps ‑‑ but decentralized production and consumption would it be a challenge making PEG, for example, unhappy or ‑‑

>> Unhappy and uncomfortable.  Well, you use these words, and I wouldn't use it.  Well, we are on our transition path and transition is never easy.  Transition is never easy.  It's always how you said it before, a challenge.  It's not around adversity or a problem.  It's a change, but we are committed to bring transition.  We want to become a green company, so we cannot assess it as uncomfortable or how you said "unhappy requested about it.

>> Right.

>> It's not that kind of question, and also we are committed to the climate policy.

>> So it's important for the whole world and also for Poland nowadays to decarbonized but as I said previously, it's ‑‑ there's so much importance into doing this not emotionally, not, you know, without ‑‑ without thinking.  We have to do this reasonably.  We have to do this, and we cannot do this ideologically.  We have to be smart about it.  We have to be smart about it and have a roadmap and ‑‑ we will as Poland, yes, we need roadmap, which will be kind of tailored solution for us.  It's obvious ‑‑ it's obvious because we are coal ‑‑ a coal‑reliant nowadays.  We have to have a tailored solution.

>> Right.  So let's try to draw this path ‑‑ thank you very much.

     Let's try to draw this path and starting drawing it from the technological point of view.  Let me turn to Professor Juan Garcia Guirao.  Probably, Professor, there's some hope in rethinking, you know, the way we use and the way we produce energy, and it seems that this path would require, you know, endless number of sources and equally large number of users, consumers of energy, at what stage is science now?  What are most promising projects in Spain and in Europe regarding modeling and optimization of energy green managements?

>> Again, first of all, thank you very much to the organizer for inviting me to take part in this forum.  It is not common to have two mathematicians in this environment, so I am gland the international scientific committee realize mathematics modeling is going to play a future role in the future strategy of energy management.  I mean, the global management of emission reduction.

     Let me talk what I need firsthand, which is the Spanish situation.  In the Spanish authorities exist of identifying as a central problem the optimization of the energy management and as an example of this as you can see they did buy the Spanish government and the rules to the benefits of the two companies.  You can imagine the big troubles that are generating two families and companies due to the revenue scale of the electricity price where prices more than 200 Euros.  That fact has made the increase of the influx with the reduction of the consumption of power and with the increase of price of all manufactured products, which is linked to the increase of the production costs you know, so the increase of this electricity price is connected with the petrol increase and which makes transportation more expensive and in short the final damage are families and companies and many of these companies are forced to close in Spain after almost fighting gains of the epidemic of the pandemics, unfortunately.

     In this frame the minister of the government of Spain and the offices of the communities have supported many projects to increase the optimization of the energy, pushing the green and clean energies.  And, again, here let me talk my personal experience on it.

     I have elected as a director of the chair 4.0 in the frame of the bilateral relationships between Primafrio and technical University of Cartagena.  Primafrio is the biggest European logistic group devoted to the transportation of refrigerated goods and tracked by route.  I'm pretty sure that you have seen on the Polish highways the Primafrio because we a fleet of 2,500 trucks because they are distributing fresh products in all countries.  They have reached an agreement with the University Cartagena where I work to build leaders in the so‑called green logic and the main project that they I am helping to push in the frame of this country is to join a big European consortium participated by big companies like Mercedes, Volvo to make real truck.  You know there are some prototypes of particular trucks that are working with combustion engines, and now the new challenge is to transfer this technology to trucks working along and routes with solar‑assisted enabled refrigerated trailers to make such transportation of this fresh product 100 percent clean.

     You know that the technical problem that this project has in the truck is something that ‑‑ is a problem that the community are working on that there is another big problem which it is generation of conservation and distribution, and there's a big consortium here in Spain leading by companies and participated by Primafrio to design a production distribution here in Spain, and we have the compromise of the Spanish authority a significant part of the next‑gen to push this project in order to try to make it a real ‑‑ a real fact. 

>> Is there ‑‑ is there a role for AI in those projects? 

>> Excuse me?

>> Is there a role for AI in those projects, artificial intelligence?

>> Of course, of course, of course.  In this project mathematical models and artificial intelligence will play a central role because you know these trucks need many, many components to make real the development and, of course, yes, artificial intelligence will play an essential role, and there is in the consortium more than 25 universities with their mathematical departments trying to push this project, yes, of course.

>> All right.  Thank you so much.

     Let me turn to the professor with a similar question really.  Any promising projects in Poland and perhaps internationally that you could mention in mathematical modeling as a way to manage challenges of our transitioning energy markets, which is transitioning much faster in our goals, in our regulations but not as far as, I'm afraid, in terms of technologies? 

>> Thank you very much for this question.

     So I would try to answer it as a mathematician from the point of view of mathematician.

      Today there was the role of artificial intelligence so putting these into context, so the reason here we developed a huge development in artificial intelligence, new models of artificial intelligence numbers optimization methods, and we can observe it in life in us, in cell phones, everything that's smart ‑‑ there's applications helping you and, of course, we also would like to use this in energy management so let me just give you some ideas what companies I cooperate with, what to do and how to improve energy management.

      It was also mentioned today that there is high investment into renewable energy sources and nowadays it also happens that the level of end users ‑‑ so if you look around in Poland there are many houses with photovoltaic panels so users can have it at home sort of can become producers and consumers, and, of course, it goes to thousands of people and that produce ‑‑ on one hand, it's very nice.  But on the other hand, it generates problems because if you look at two houses, there's some families, but they have different habits and behave differently although the installations look similar, and, of course, it's hard to, like, describe it by particular mathematical model, but it's where artificial intelligence comes to help because these models of artificial intelligence ‑‑ they have property they can adapt to patterns, they can recognize personal habits and that can be used to make effective management of energy because you can ‑‑ you can put if in every house and collect the data and then make model of village and then bigger pattern so on and so on.

     The bigger problem ‑‑ to do that you need lots of data and nowadays you need to collect data but now you need infrastructure to get this information.  You need mathematicians to tell you which information is important, which is not, how to fit these devices, these artificial intelligence devices because, of course, you cannot put everything because that way you will lose ‑‑ you will have too much data, too much competition and power you cannot do that.  So for that you really need clever people to think how to ‑‑ how to ‑‑ how to define models.  What is important, do some statistics and prepare everything, so this is really ‑‑ really challenging problems.  But if successful, they can, of course, help in better information because the more information you have then maybe you can forecast better or you can know how much energy you should store and how much you should produce and that really believing help in management.

     The main problem is that you need digital ‑‑ digitalization because without collecting data, you cannot ‑‑ you cannot make such decisions, so you need really good infrastructure.  And if you have that, then it really can improve ‑‑ and if you, of course, manage energy storage better, it saves energy, and that's really good. 

     So these are really challenging problems, but people work on that nowadays.  Thank you.

>> Just out of pure curiosity, how many years of history of data collection we need to be able to correctly restrict the usage and production of such distributed production and consumption?  Meaning that one year is enough and next is similar or you need 10 years of history of data to be able to predict what will happen in the next year depending on, you know, how the weather patterns change and temperatures change and loads of people ‑‑ you know, can you give us a feeling of, you know, what data ‑‑ what informations would need to be gathered ‑‑ and I understand are not yet ‑‑ at least not in sufficient way to be able to really say, you know, we have this big data model which tells us what will be happening next 'cause forecasting is just, you know, telling what will happen in the future and no one ‑‑ nobody really knows or maybe mathematicians do?

>> Yes.  Much depends, first of all, on how much data you have because you can collect it every minute or maybe every day or maybe every month, so, of course, more is better, and you should really try to do that as often as you can and then also depends on how much you want to predict if you want to know what will happen tomorrow.  Maybe for that maybe one week will be enough.  But if you want to predict one year forward, then you really need lots of data; right?  Because you need weather ‑‑ you have to take into account many manufacturers, and, of course, that long prediction won't be accurate.  But in the short‑term, it can be quite accurate, and you can have really, really good results like maybe two weeks, three weeks.  That can give you a satisfactory accuracy, of course, you have to collect really in a good manner, and, of course, you have to prepare your motors very well.  That's the main ingredient because not collect huge amounts of data and look into it.  But you also have to trend your motors:  What is important?  And then you have a chance with results.

>> Thank you so much.  That was ‑‑ that was really interesting and I ‑‑ to those of you that think all IGF is about internet really and everything that's ‑‑ that surrounds ‑‑ well, I don't know, surrounds internet or internet surrounds other things; I'm not entirely sure how we should put it.

>> Excuse me, if I may step in.

>> Absolutely.

>> The question you've asked, because you've asked ‑‑ you said that it depends how many data are we do collect, and so, first of all, we need ‑‑ I should point out that we, as an energy company, are provided ‑‑ are obliged to ‑‑ to soon to measure every 15 minutes the ‑‑ the ‑‑ the energy consumption with our clients so ‑‑ and to do so, we have to start from 2023 to change the ‑‑ to change devices, and we will have time till 2027 to change all devices ‑‑ to change devices ‑‑ to all our clients, so all of them ‑‑ so we would be able to collect the data every and energy consumption every 15 minutes from every one of them, so we will have a lot of data, but also what we need to remember, and I think we said this in this panel discussion ‑‑ it's not going to be a linear growth because the consumption is getting different, much more different than it was in the past, so ‑‑ so I think that's why this also is the area where the ‑‑ where the artificial intelligence will be applicable and will be necessary and will be necessary so much, so I'll stop.

>> And I understand there will be a huge need for this data, and I would like to thank you very much, Mrs. President, for stepping in because I was just writing down that in the end of this series of questions, I will ask you about this metering thing because it has been discussed in Poland for quite a long time, and we still don't have too many smart‑metering devices.  But, obviously, there is a growing need for that and not just for creating dynamic charging for energy pace and to collect data and monitor it.

     Let me turn to Professor Aidan O'Sullivan, who's a specialist in AI applications in power generation. 

     For our first question ‑‑ if you can tell us how or in what way artificial intelligence, AI, can be applied in energy generation?

>> Thank you very much for the invitation to be here; also for a very straightforward question, so it's my favorite type of question.

      In terms of how AI can be applied in the energy system ‑‑ if you kind of zoom out, one of the challenges within the energy sector is that we have this commodity electricity.  It's a very unique, unusual commodity in that it has to be consumed as soon as it's supplied ‑‑ you know, gold doesn't obey this rule; oil doesn't obey this rule.  If you don't want to sell somebody oil, you can put it in a barrel and put it in a room but electricity has been to be consumed right away.  It's one of the things in the electricity sector, and the way in which ‑‑ the way in which artificial intelligence can help is that it has ‑‑ if you look at it broadly the characteristics of a software technology as opposed to the renewables, which harbor technologies and so forth. 

     So while we're introducing a great deal of complexity into the electricity system by adding renewables, artificial intelligence has the characteristics needed in order to compensate for this degree of complexity, artificial intelligence does really well in problems where there's a great deal of complexity in the power sector, we don't know what the correct action to take is or the correct ‑‑ the most optimal performance it is just because the scenario we're working in is so complicated. 

     You know, three‑phase power and moving electricity around the grid and scheduling generation, these are all highly complicated problems that we sought to solve in a kind of proximate fashion, you know, relying on markets, relying on operators kind of know how over years ‑‑ however, it's always been challenging to say what the optimal action is

      However with the artificial intelligence, you have the capabilities to say what the optional optimal is and to take that action more on a regular basis, so this is how we can see gains in flooring efficiency through the application of artificial intelligence by taking more optimal actions.

     And one of the best kind of research projects illustrating this is in Europe right now iss the learning to run a power network challenge, which has been run by the French National Grid RTE in collaboration with the Electric Power Research Institute and ourselves at UCL as well as some at Google grain and other industries, so this was a challenge accepted, which is the largest AI conference in the world to allow teams to compete, train artificial intelligence agents, to manage a simulated power grid under disturbance.

     So the different teams competed to see how long they could keep the grid up without experiencing a blackout while renewables ‑‑ while lines were closed and cascading failures and things like this so a really exciting piece of research done.  It's a very open and collaborative way by allowing innovations to emerge from universities all over the world and the eventual provided in 2020, so it was great to see big tech taking a part in complex power systems. 

     So the context of artificial intelligence lens itself is very well to some of the problems we have in the technical sector to the electrical system, which is designed for residents.

>> How far from research we are to actually allowing AI to control the grid, the system?  Do we have enough data, data points, measurements that would allow the machine to know what is actually going on?  Is it ‑‑ you know, are we like years from the moment that we can have full automatic control if I'm calling it correctly, in respect to artificial intelligence or we're more or less, you know, technologically ready already?

>> Yeah, it's a really good question, and I think it comes down to an element of trust.  You want to build up trust in these systems, so I imagine the first way that this would evolve would be as a decision support system for operators giving them a new tool to help them do their job as we're making it, you know, considerably harder.  It seems only fair to offer them a new tool to help carry the load a little bit.

     Could we switch to kind of full ‑‑ already control there would certainly be advantages in decision times and capabilities that far exceeds the human capabilities being encouraged by this; however, from a regulatory kind of perspective, you could see there being challenges around that.  So while I think automated control is certainly possible and in some sense a circuit breaker in a grid is already automated control, there's not some human standing over that examining the conditions because it happens so fast it has to be automated.  And in some sense, you know, it's the same argument with autonomous vehicles is it too dangerous to allow humans to drive cars when they can be done more safely by AI agents?  Maybe the same question will be asked of grids, you know, is it efficient to allow humans to try to control something that's been described as the most complex machine ever built, you know, grids span thousands of miles and connect into each other and, you know, a myriad of just ‑‑ a miracle of complexity science.

     So in one sense we might look back in 20 years and kind of marvel at the fact human operators were able to operate grids so stably and maybe we need the two to hit levels of renewable penetrations of 80%, 90% that we need to mitigate climate change. 

>> All right.  Thank you so much. 

     We're not sure ‑‑ is Mrs. Kendra Pierce with Microsoft with us? 

     You remember in the beginning of our conversation, I was muting and unmuting myself trying to figure out if there's one more participant, but I'm afraid not. 

      The question that arises when we think of all the computing power needed to control the grid and also to the system ‑‑ I don't know, maybe a politic chain system of dealing with paying for all those things by consumers to prosumers and to professional producers is the amount of energy that might be used by this technology ‑‑ unfortunately, we don't seem to have someone who would tell us more about data centers and their use of energy which also on the other hand you probably ‑‑ we'd probably all heard there are data centers, which are energy‑neutral, so they produce just as much energy as they use, and they do not produce greenhouse gases so ‑‑ well that's another dimension:  How digital technologies can help or perhaps use the energy that efficient households might save. 

     So perhaps with that I could, again, turn to Vice President Buk, you know.  Did you have any observations in PEG regarding the computing power needed for ‑‑ for those iteration.

>> I would say there's a necessity to start with full replacement of the electric energy meters; right? 

     And I mentioned already that we are at ‑‑ well, at the very beginning of the process of the replacement because ‑‑ because till 2023, we ‑‑ well, the whole Poland system of 2023 is obliged to replace 15% of meters with the new devices with the new functionalities.  And in 2027, the Polish energy companies are supposed to replace 65% of devices, so we have to start there, first of all. 

     And then there's another issue that we have to that we have to bear in mind, does this ‑‑ is this the adjustment of the distribution network?  Because nowadays we all know that our distribution networks are only adjusted to deliver the energy to our clients and in the future, in the near future because of the growth ‑‑ we already observed that, right, the growing share of the ‑‑ of the consumers.  There will be ‑‑ the energy networks need to be adjusted, so they could also receive the energy so, you know, you are thinking of the competent power.  And from my perspective there is more significant challenge in the whole replacement process and in the adjusting ‑‑ adjustment of distribution network because this is ‑‑ this is an investment process, which will ‑‑ which will take time; consume a lot of money, of course. 

     Also we will take advantage from this in the future once we ‑‑ once we will be able to provide the best ‑‑ well, the energy efficiency in the best possible way but still, still it will ‑‑ mostly, you know, adjustment of the distribution network will take ‑‑ will take some time, and this is what we need to do ‑‑ what we are focusing on at least from our distribution company.

>> I'd like to comment on energy consumption and data energy and components of AI in that space.  So while I'm based in the U.K. I'm actually Irish data centers have been a topic of conversation with limits being placed on the energy consumption that they can draw at peak times and wanting to keep in mind while training artificial intelligence algorithms is that they can be very expensive and require large amounts of energy including an GP3 large language model which involved Microsoft as well, you know, trained for five weeks in a data center consuming millions of pounds' worth of energy.

     What's important to keep in mind however ‑‑ there is this, you know, benefit of scale that you get with a data center and if you try and do that locally the experience would have been even more expensive die to the lack of hardware, so there's a reason we have data centers.  They do kind of come with benefits of the scale of computation that they enablement for these data models to be trained, however, it's important to be mitigate the computation that comes with them.  And one way to do that is make sure the models that are developed there are used. 

     Again, if a model is trained once and never used, it's a complete waste of energy; however, a model that's trained innumerable times ‑‑ a model trained once and used innumerable times you offset that by usage and being intelligent with it and making sure the investment in the energy consumed by the data center results in tangible benefits is really the critical kind of energy that needs to be taken into account.

>> I would very much encourage every participant to step in so ‑‑ it's just hard to manage those actions, but you're doing ‑‑ I mean, almost like we were sitting in one room, so I'm really grateful for that because this exchange of ideas between panelists is key of discussion.

     Mr. Professor, would you like to add something?

>> We spoke about these energy meters that should be replaced, but, of course, we can go a step farther because if we have, like, a big building with offices, we can ‑‑ nowadays we can measure everything like what its usages of energy from circuits?  What's the usage of energy from air conditioning, frozen devices?  We can measure all of that, but, of course, it's a matter how deep we should go and how much it can help because it can happen that maybe it's enough to measure at the end of the line and maybe that's enough for prediction or maybe we should use the thousands of sensors and then try to predict so there are definitely ‑‑ there are opportunities like that and that, of course, will generate a huge amount of data and will be ‑‑ and everybody is ‑‑ because putting some sensors is the effort and maybe that can help, so that's, like, step further compared to what's a governmental level, and that's also for the future that if you have big office, maybe you can optimize your consumption and make some decision with company that deliver energy.  Thank you. 

>> Thank you. 

     And, of course, 5G and IoT are also coming to help because they will naturally enrich our lives with numbers of different measurements and meters so with ‑‑ and meters with smart home being probably a good source of information about ‑‑ at least this more affluent households, how they use, when, and this will contribute.

     You know, we have 13 minutes left in the room here.  There is ‑‑ there are not so many participants but there are some.  I think I would be very happy as moderator if we can involve also participants, so I don't know ‑‑ if anyone would have any questions, you can approach the table and ask a question.

     I don't see much enthusiasm.  I don't think we can ask participants digitally to ask questions.  Can we?  This is me looking at our control room, not really.


>> No, we can't.  Maybe we can slowly start wrapping up and ‑‑ you know, the Vice President Zuk asked not to be too emotional in this discussion.  I would like to ask you also for honest and maybe even emotional picture that you could draw of AI if you have the energy.  Because if we look at different trends, we might be afraid that we have to scale down and greenhouse gas emissions very fast.  We don't have sources of power for huge generation, professional generations like we do in nuclear power plants or coal power plants.  So at least myself ‑‑ I can only see the future in small distributed grid where you have hundreds or maybe thousands of small energy producers, not just prosumers but on the little rivers ‑‑ little windmills, solar power and all those things consumed locally because energy is the commodity that cannot be sold later as Professor O'Sullivan very well reminded us.  We cannot store it in a barrel electricity.

>> But what you're saying ‑‑ this is that you imagined that in the near future there will be some local power generation plants but small plants or small wind farms or something like that that will lead to energy consumption locally.  You know, it's not how it works right now ‑‑ how it works right now, and it's not how it will work soon because as I told you ‑‑ as I already mentioned ‑‑ as I already mentioned in this panel discussion that it's also about the distribution level that needs to be adjusted, so, you know, there is a lot of challenges ahead of energy companies, again, ahead of distribution companies also and for our guest ‑‑ it's a Polish ‑‑ let's translate Polish energy network's company ‑‑ it's like the main one which ‑‑ which manage the distribution as a whole, so these are huge challenges ahead of them to adjust the networks, to adjust the whole system of networks and renewables, to more decentralized system of generation ‑‑ of power generation, so ‑‑ well I would rather ‑‑ and I ‑‑ I'm not sharing ‑‑ second‑guessing your vision.  I would rather ‑‑ I would rather expect ‑‑ I would rather expect sources based on nuclear power as a backup what I already mentioned because we can see in the whole world that this, you know, nuclear power ‑‑ nuclear power is on the cusp.  We can observe it, and it's not deniable.  So from my perspective, this is how the situation will look like in the energy sector soon, in the near future and, and still I find the need of being ‑‑ of being professional being extremely here and not emotional, and I know we can ‑‑ I'm aware, and I feel the need ‑‑ I feel the need ‑‑ I share the need of the decarbonization, but still there's a whole society ‑‑ the whole world and the whole society and the whole economy where the energy's needed, so I cannot imagine the blackout ‑‑ the acceptance for a blackout ‑‑

>> Can you ‑‑ can you imagine an offgrid household?  I'll ask our mathematicians on the panel if you can model this sort of ‑‑ I don't know one or maybe a cluster of households offgrid

>> I can imagine that for sure because technology can provide these solutions, you know ‑‑ , but these kinds of solutions needs time to be implemented so ‑‑ so, yeah, but, you know, the possibility that will ‑‑ that we can benefit is one thing ‑‑ is one thing, but the time when we ‑‑ for income and patience for these kinds of innovations is another thing.

>> All right.  So let me ‑‑ Professor, you're sitting across the table from me, so it's easiest ‑‑ can you model for me my household as an offgrid or maybe me and my neighbors?  Can we do that?  Is it model‑able.

>> So let me answer your question so, of course, mathematics is in service of technology.  We cannot, like, you know, make something artificially.  We need data, and then we can help, but, of course.  As a mathematician, I cannot build infrastructure.  I can tell you how you should make your infrastructure effective or what you ‑‑ what you should to work it better, but, of course, we won't create the infrastructure, so that's how we service which helps to make smart decisions, but we cannot do more.

>> So anybody else would like to ‑‑

>> I agree with Professor abroha, mathematics these days can help plan a future ‑‑ a future strategy.  But in Spain, the same problem that the vice president has pointed out about the distribution net is on the table.  We have to adjust our ‑‑ our distribution net for the producer because the management of the distributions have been a big issue so in this ‑‑ in this case I think that mathematics are two or three steps forward to the ‑‑ to the implementation of the technology that our network is needed, but I'm pretty sure that in the next year we will ‑‑ we will do the necessary step for ‑‑ for it because we need to change our model; make clean everything and to ‑‑ to forget carbonization, and this is ‑‑ this is ‑‑ this is our future challenge. 

>> Thank you so much.

     Change our model that's a quote that's very strong, and I think we should remember about it. 

     Professor O'Sullivan, would you like to add something to my question about offgrid?

>> Yeah, I would probably disagree with that.  I think there are a lot of benefits to being on grid and the benefit is scale, you know, as we need to move quickly.  There is this issue around energy, equity and the costs and expense of the transition.  It's better for everyone to have, you know, it distributed across a larger amount of people, and that kind of works better rather than say wealthy households buying solar panels and wind turbines in the becoming and then leaving the cost of grid maintenance to the people who can't afford to buy these technologies so, you know, things like the Tesla power wall and all that are really good, but they're targeted at a certain sector as are electric vehicles, so the scale of the challenge means that its more advantageous to have a setup where it's available to all consumers in the system.  And, again, energy equity is a key kind of challenge in this ‑‑ in this trilemma that we have.

>> So let me use the last four minutes of this discussion to ask each of you to tell me in 10 years from now, hopefully, we will not be living in a desert yet but hopefully we will be on the way to avoid the serious and climate catastrophe, but how will the energy center work in a very, very brief and concise way each of you having, like, 30 seconds for a short, you know, picture

      Let me start with Professor Aidan, you were last speaking.

>> Sure, I'll try and be quick.

     I think when we look back at 10 years, we look up the current state as the real desert, you know, what we're doing right now is we're building the digital energy system, which is a representation of the energy sector and digital space that far exceeds what we currently have, and that will be a much richer representation that will enable all types of applications and the interconnection of energy and transport of interconnected energy, and I think the future of energy is incredibly exciting, and it's a really great place for people to work, especially young people, with a technical background.

>> Thank you very much.  Professor.

>> I'm sure it will be different because now everything is changing so much so.  Definitely I expect that artificial intelligence and digitization will ‑‑ will take huge ‑‑ huge impact on change and yeah, this will be definitely much different than we have now, and I hope much better.

>> Professor?

>> I'm pretty sure it will be much better because as a global community, we have identified this problem as a central problem, and we are all together working to more clean and efficient energy. 

     And here in Spain you know we are a real desert at least in the middle part of the country because of the weather.  And if you take a look to our small town halls, panels, solar are almost in all the houses so there is a real ‑‑ a real aim of changing and produce clean energy much better than now for sure.

>> Thank you so much. 

     And the last words belong to Vice President Buk because you are the one who will build this different ‑‑ or your company rather together with you, hopefully, will build this new grid, this new energy system.

>> Yes.

>> So how will it look in 10 years?  I understand you will never agree to realize where poor families will be cut off because they will not afford so, you know, how ‑‑

>> Your vision ‑‑ the vision that you mentioned and also I second what the professor said because it's also the question about the energy efficiency; right?  You know, local consumption of the ‑‑ of the energy production, and so but answer tog your initial question I'm pretty sure we will be like Poland and the Polish energy company will be a less coal‑reliant country, a less coal‑reliant company.  We'll still be on our transition paths, so it will be ‑‑ it will still be in the, you know, in the ongoing process of the transition and building new capacities, new green capacities will be more digitized for sure and will be more energy‑efficient.  That's the future in the next 10 years but still we will be on our ‑‑ on our way to achieve net zero.

>> Thank you so much.  It's been a very interesting discussion.  My voting did not work.  I'm terribly sorry for that.  Nevertheless, you know, I hope we're going to a future where there will be no desert and also no darkness, meaning that we'll overcome the challenges.

     After this discussion, I'm not entirely sure that we know how we will do it, but I'm certain digital technologies, AI and mathematicians working with it will do a lot to help achieve our goals.  Thank you so much, and I wish you all the best for the rest of the discussions here in Katowice until the end of the week or until the 10th of the week.  Thank you so much.

>> Thank you.