Blockchain and AI to Support ESG Agenda: Latest Trends & Use-Cases

Transcript

well thank you so much for being here again some great panels that we've had uh so far another great one that we're about to have uh by our lovely moderator Maya here who's going to be talking everything blockchain um Ai and sustainability ESG fantastic like take it away let's give a round of applause thank you uh do we get that one as well yes okay so I'll start um so as U it was announced now it's time to talk about Ai blockchain and what Ai and blockchain can do for esgs and to support ESG agenda I am actually super happy to have this group of people sitting with me today and aside of them being fantastic personalities um I'm happy because we

are actually covering the entire scope of esgs starting from uh sourcing the energy itself to data uh green washing and regulation uh so we have everything covered so without further uh Ado we have Mario Gemma uh Mel and Michael I'm not going to do any surnames because I'm going to butcher some of them uh because we have actually Italian mafia here I'm just I'm moderating I come in peace so um guys please few minutes just to introduce yourself uh and yeah projects that you're coming from and then we'll just go straight into the topic thank you um my name is mar Alberto Kazi I'm the CFO at sing net and co-founder at

Singularity Dao uh Singularity net is um the oldest uh one of the oldest Crypton AI project we launched in 2017 to build a decentralized network of AI applications and AI agents um and uh more recently we have uh we are uh undergoing now um uh merger with two other large crypto project fet and ocean protocol to form the as I super Alliance super intelligence Alliance um and um to Foster the future of the centralized AI we believe that AI should be decentralized um in the contributions in

how it's run and open source so obviously with the rise of llms uh things are uh you know exponential growth has come and uh it's the right time to push for it to be open and for the benefit of the uh why they're public thanks uh my name is Jemma green I'm co-founder and chairman of power Ledger we make software for tracking and trading electricity and trading uh environmental Commodities like renewable energy certificates and carbon credits we've been operating for eight eight years we run uh public blockchain we created a public blockchain called the power Ledger chain which was originally

based on ethereum which we ran as as a private Consortium chain and we um made a decision to hard Fork Salina about 2 years ago and adapted that further for the energy use case and made that public uh we have our own token the power token which is um used to pay transaction fees to write data to the blockchain and and our clients are utilities and large corporates that want to track electricity for procurement and Reporting purposes and utilities that want to make new products such as peer-to-peer trading of electricity new commercial models and new business models and uh we were an Australian company but nine months ago we moved the company to Switzerland along with uh my family and my business partner and his

family as well thank youa hi I am Mikel suavi Chief sustainability officer at impas cope where we leverage web three and artificial intelligence for sustainable solution and we have uh two main projects we are working on one is a decentralized Marketplace for impact with a dow component and another one which is already up and running is a Ai and web three based green washing identifier and uh on top of this I want to mention as well the fact I'm member together with Maya of the sustainability working group at Crypt valy Association

where for the last year basically we have been studying the intersection of of blockchain AI and sustainability last year we published a report which you find on the internet site if you're interested and yeah the development is going on therefore keep an eye out for the site for more to come thank you all uh my name is Michael brly a director Ai and partners what we do is we're a Professional Services firm for company subject to the EU AI act for those of you in the audience may know it's a new regulation of AI coming into Force this month or next and we provide reg tech software advisory and consultancy services the inspiration

behind this was gdpr implementation so if you had the regulation of data this is now the regulation of AI so the analogous equivalent is there to the extent it touches sustainability I help write the standards uh which high risk AI systems have to be conformed against for example to meet certain ESG criteria sustainability and address the consumption issues which the panel probably going to touch on um also advocate of the work that Michel and um crypto valy Association do and it's a pleasure to be here with you all thank you everyone um so in case someone was living under the rock lately uh ESG is um econ um environmental social government uh there is a ongoing

narrative is it agenda is it criteria whatever but we know for sure that uh reporting is uh coming in place and uh there is this entire narrative about investment in esgs and so on but to begin with environmental part uh I don't think we can talk about it without electricity like resources for for for the energy part especially like in in AI Factor because AI is like on Hot Seat uh regarding the energy usage and so on so how de centralized and peer-to-peer uh um energy trading can support esgs to begin with it's like a

source well I mean in terms of electricity you know it pretty much remained unchanged for like a 100 years and then about you know 30 years ago we started to put like distributed renewable energy into the system but it was still run by like a centralized market and the idea of decentralized and distributed energy is about creating local energy markets so people can share energy and trade it locally which is um you know a nice thing that people like to be able to do um you share energy with your neighbors give energy to your local school but it actually has a functional purpose that helps the grid as well because um if we all live on the same street and um you have a battery I have solar and um I'm not consuming all

of my solar I store it in your battery and buy it back at night which I like but the grid also likes that because otherwise that electricity is being exported and creating congestion in the physical infrastructure and if we actually work um create the the market dynamics together we can relieve grid congestion and avoid having to upgrade infrastructure like um substations and Transformers and capacitor Banks and then I think the other opportunity around decentralized compute and decentralized electricity is really co-optimizing those for low carbon outcomes low cost outcomes as well so I think there's going to be a really strong intimate link um on in that regard so when I think about like ESG

sng feels more like a cost center whereas energy feels like um you know doing good but also doing well like it's really like a huge business and the amount like people talk about energy transition and I think that's completely the wrong phrase we're really in a state of energy expansion and it's actually always been that way every time there was like a new energy source from like coal you know to oil oil to gas um uh nuclear Etc we just added those energy sources on top of all the other exist pre-existing energy none of those disappeared and so Renewables is just a new addition into this EOS energy ecosystem and we're actually consuming ever more amounts of energy and you can

see from The Grow trajectory of AI that that's going to become like a new energy hungry sector that's um you know we're only just at the beginning and um yeah it's it's going to be pretty profound well I I guess the decentralizing very traditionally centralized environment is extremely challenging like how much support do you get from yeah I mean I've kind of what I just said now makes it sound so easy and I think thanks for kind of yeah it's it's so not like uber where you know you can just create um an app and everyone can get on the road electricity is highly regulated and all of the incumbent players make money from the system that it currently is so there's a

lot of resistance and entropy like when we said the company we were doing Little Lighthouse projects in different countries based on where there was interest but what you know we've had the business eight years now we got up we were up into 15 countries operating uh demonstration projects because you know an Indian company doesn't really care that you've got a project in France they want to know if you can do one in India uh and so you ended up having like you know a lot of different demonstration projects then you look at where's the opportunity for scale and that's really a function of um with the regular fit for purpose now or is it changing and so when we started to do that assessment along with the addressable Market size then we've just focused in on actually five jurisdictions this year

and that that's Europe and India which was in part why we moved the business to Europe to be closer to the markets in which we operate um as well like I have a young family and traveling from Australia to Europe regularly yeah was not the best the most practical thing yeah exactly but I'd say that it's it's very it's like GLA it moves at a glacial Pace energy but I think things like Ai and that attention in the system it actually will be catalytic to start thinking of new ways of doing things because the cost of electricity now is so uh material now to the cost and you know we're having we're seeing things like you know offshoring of Industries because it's at a threshold point so I think that's going to necessitate more

Innovation than perhaps we've seen uh in recent years and then because you're um combining blockchain and all of that then you can trace and then uh you can have like a certificate and kind of minimize green washing for example right uh yeah that's it uh I mean happy to bring my my experience on basically trying to leverage both Technologies at the same time and just to give you an idea we created this tool where we extract you know corporate Communications basically social media financial report Su sustainability reports to generate a qualitative and quantitative report on green washing so

I mean green washing is basically the where where greenwashing is the practice of you know overstating the environmental position of your company and as there are a larger number of regulations which are coming out sustainability reporting or uh European green deal which you know they're somehow very happy heav regulation for the time being in European Union they could punish you with up to 4% of your revenues if the find doing green washing and yeah therefore with the ability of using the AI component on the one hand to generate a report and afterwards with the

possibility to save the report on the blockchain we are connecting the two words and two completely different paradigms because on the one hand you know you have the artificial intelligence the Black Box you get very nice results but you don't really know much how you got such results whereas on the other side of the spectrum you have the blockchain you know the permanent record which is there which is permanent which cannot be changed and therefore in a way we are working on creating some element of explainability on AI you know we are trying to pass from the black box to the white box from that perspective

and uh yeah I think it's uh yeah there's a lot of complementarity to be seen there between the two technologies and uh I think it it just started in a way okay and then like we can't avoid the question of Regulation I mean with with the recent regulation from EU for for the AI but I get again I really really want you to explain what you told me uh recently about AI washing which is super interesting to me but also like reporting on esgs is that being addressed um by EU and standardized in any way um so AI washing which uh is a term AI Partners came up with refers to it's

the equivalent of green washing so green washing as M stated is you're overstating the environmental benefits of our product in then AI term you're overstating its trustworthiness its security explainability it's not so you're saying you're seeing a plethora of products in the market saying we're very trustworthy we're very secure when in fact you're not so that is AI washing um how it leads into the great work there we with with reporting um so you can the AI use cases are growing so you're going to have to report on ESG and you can use AI to optimize that reporting process its capabilities to analyze large swords of data are well known to the market however there is a there is a running a risk of saying airic Victory so if

you're using AI to uh to optimize reporting any issue you're meeting that however you could um the computational cost of running an AI system could actually exceed that which exceed what you had done with the ESG reporting so in essence you're using something which is going to have a worser effect than you were originally started um so that's also an element to consider now what we talked about with the eui ACT it's a risk-based approach to regulation the best equivalent again is kyc financial crime show up hands who here has heard of kyc cool most of you so you have to basically identify your customer and risk classify it because that determines how you treat them the same thing

happens with an AI system everything from a simple chatbot to an sophisticated or um generative AI product that black cr's using for example investment managers are now using you would have to RIS classify it it's KY AI um and it ties in with all these new regulations CU any use of AI would fall within scope of this so if you used AI to comply with a separate regulation bear in mind you still have to comply with this one but uh happy to answer any questions if you guys have any on that so Mario how is merge going I'm joking another topic um no now we have um like the all those things how can we use AI or can we use Ai and all the

existing uh infrastructure to potentially uh have ai standardized for example reporting because it can pull all the information that's already on the blockchain um I think J said something very interesting Ai and uh with the manifestation of llms has has has been a catalyst also for many to start thinking about how to make things more efficient yeah data is available right uh data is available uh in uh though uh secluded places somewhat uh so you need to go and find it or you need or you need the you know those larger um entities or institutions that have that data availability to make it available

uh once the data is made available and you need to have the right incentives for people to want to process it and uh find a way to make things more efficient uh whether it's on energy or um on uh on on supply chain for examp example right so I think that uh with what's happening right now this has been a catalyst for data to be more open by way of wanting to find a solution to things uh and obviously AI by uh the by by by design by the ability of solving problems is able to find uh is able to find Solutions uh is able to make things more

efficient now we're just at the beginning of this because uh llms are a very specific application which is the one that is getting the most adopted right now but is text based uh so there's some constraints in the analytical part um where if you're um more numbers or you know uh algorithmic base you're still on reinforcement reinforcement learning methods which are the old ones um it also uh it's interesting because on one side it's a solution on the other side it creates further problems uh or it may because uh it's very uh very energy energy consumption is really high right so

we've just starting now and uh uh uh the need for compute is almost infinite it's probably starting a super cycle that is going to last a decade and more of continuous increase of demand for compute and compute needs power and needs a lot of it uh and so uh we could use AI we can use Ai and things like uh Power ledger to make energy usage and consumption more efficient but also what we want to do is make the use of the hardware more efficient because uh right now the base model is heavily

centralized right so we're going to servers the big players right the big TXS uh but the you usage uh of uh of the capacity is not hyper efficient like often times that Hardware is unutilized and so that's one of the things that uh blockchain uh can help with so with the rise of distribute compute networks for example you can always go and uh tap into unused and idle power uh which is a promise that has been there for a while people have tried already in back in 2017 2018 uh but now it's a major Trend and it's one that is becoming increasingly important

because now the demand is actually there the supply is actually there and the demand is actually there so creating the C creating marketplaces for AI compute like market prices for uh energy to try to make the supply chain more efficient uh is something that is highly valuable for sure thank you uh JMA something that um you mentioned you worked in the like the use case um so does it tackle somehow the social part um like I I think it does um yeah in a like if you think about peer-to-peer trading if I can only Supply myself with energy with my rooftop solar I'll put a

certain size system in but I might have a big roof and if there's a market mechanism for me to sell it to you then you know that's allowing me to you know basically monetize my roof space and generate more income for myself which is a social benefit as well as like an environmental benefit in local energy renewable energy low carbon or zero carbon energy me on the other hand the choice exactly and you might be renting and you are not allowed to put solar on your roof so you can still buy renewable energy locally from your neighbors as well so I think there's definitely um like many social benefits associated with uh Ren local Renewable Power um one question that I personally

been thinking about is so ESG is now super popular even though it's not the term that's like two years old but now this entire thing about ratings when like companies initiatives everything is rated based on esgs how do we avoid um not maybe grain washing but avoid the scenario that we had with uh carbon credits uh how do we keep the legitimacy of it and is there like maybe blockchain a I Can Lift us to the next level yeah I mean I have a bit of experience in this because I used to work in banking in ESG like looking at companies and um looking at ratings that other companies prepared for them and I I felt they had so many

problems with uh like using those in any kind of meaningful way to make decisions uh for example like Banks would typically think about like an oil and gas producer in Asia different to one in North America in terms of the environmental and social standards that you'd expect and so you got to take account of localization issues or uh you might make different trade-offs between like social outcomes when the environmental standards are as high because it's going to provide Economic Development opportunities not withstanding perhaps the environmental standards are not the same as um what you might see in high income oecd countries so I personally see like a lot of issues with these rating standards it's a malange of

factors that tries to give you one answer but there are so many different things that are important to a situation a bank or an organization that's using them so I I think that they're quite fraud and I'm not sure AI will solve all of that but I'm interested to hear from the other panelists uh yeah if I can jump on it you know sustainable Finance I think that's an interesting business you know from the perspective of sustainability basically now most of the investment are passing to asset managers basically and 50% of them they claim to be sustainable they claim to be ESG investing they claim to be impact

investing you know with different objectives and uh even though I there's not a lack you know of data or indicators the problem is that very different ESG indicators provide very different results and uh so it's difficult to understand you know which one is right which one is wrong on the other hand I think it's good that there is some variety rating because we understood we all know you know F before the financial crisis all the companies which had AAA rating which finally were far from perfect therefore it's good to have a bit of variety but on the other hand

personally the feeling is all this data World Is Still somehow too detached from The Real World and I think it's somehow still connecting to the question okay you have data on blockchain and that's really nice up to the moment when the data is digital or full that can may be made fully digital but on the other hand I would say still a problem to make sure that what is making what is happening in the physical world is properly reflected in the blockchain world now there are machine learning applications for example on GPS to study reforestation Forest coverage and therefore we can create a better

digital twin you know between the real world and the blockchain world but yeah nevertheless that's uh there must be a lot of discussion and that's somehow surprising to hear you know a sustainable Finance conference is people talking about biodiversity ratings you know luckily the conversation is not anymore just on co2 but yeah start to talk about biodiversity but the feeling is it's a very profitable and nice business but the actual impact is having it's a questionable one think Michael do you have anything to add like from the regulation

perspective basically can we hope on like real standardization can we hope for I think it's I think it's achievable so as I mentioned at the start the role of Standards is going to be huge in the eui act and ISO i e s senc they're all going to be commission by the European commission to build standards that highrisk AI systems have to conform against so for example if you're building llms and they exceed a certain threshold for energy consumption it's likely you won't be able to deploy that on the market uh equally there's also there's also a few initiatives which I've seen which are not necessarily enshrined in regulation but also have endorsement for example Stanford have developed something called the

foundation model transparency index which shows you how transparent a foundation model is against a plethora of metrics so it's likely that we'll probably see these emerge in sustainability world to rate AI Solutions on a benchmark to show how sustainable they are the same thing happened in the crypto world where I think the University of Cambridge actually did a study for Bitcoin to show the electricity consumption so I wouldn't be surprised if you're going to see the same instruments arise in the AI world because we're at the very Inception of a transformational change for Humanity and we need robust sustainability mechanisms to support a new AI economy thank you um Mario basically

this you you kind of opened the topic of uh AI is developing and it's like the need for compute and for the resources is not going to stop top so basically what we are seeing now is like next um Industrial Revolution whatsoever so how do you see this going cuz on one hand as an AI project or whatever you really need resources to to develop on the other hand there is this sustainability uh side like how how to meet those to uh well um it is a good starting point to do it uh with uh um in a decentralized setting right so as we identified um uh already Distributing networks in different ways at least create a

framework in which you can try to make things more efficient um and then I mean similarly to what we have to do with government agencies and universities and uh The Wider The Wider World also in AI it's really important to start thinking about and speaking about ethics and um how how to use it how not to use it uh and that's true for the application side as well as for the infrastructure side obviously it is a tectonic shift and it's moving really really fast you know the amount of investment that is going into AI is not stopping uh the the the the the the speed of the pace of of of

innovation is is increasing so uh they have to go in parallel because you're going to we're going to have to catch to catch it as it as it runs um so from our perspective if again we're doing it in a way that uh you know uh we really want it to be open uh and we Foster these discussions we participate to these discussions uh and being in crypto or and and leveraging blockchain we're also uh exploring for example reputation systems uh that could be a fluid way of utilizing some of the uh metrics and uh rating systems that we um that we can come up with uh but again in a dynamic

way so not you know static right so it's also about the contribution to a certain Network or the contribution to a certain framework uh that can be monetary or non-monetary which is really important as well and it can also be negative so you can have negative incentives so that's one very important thing that uh tied to incentive systems can create positive uh positive cycles in which uh you know communities of experts uh uh alliances of big players small players together can look for uh these Solutions yeah obviously like needs to be done as a as a joint venture um our time is kind of ending

but um anyone has any questions raise your hand uh use the opportunity yes please yes web I identity plays out any important importance in this workflow and on this I know that Michael mentioned the KY of kyc but then I I wonder to what extent kyc may or may not some privacy concerns the two

questions if I can Rec reply quickly on my part I feel like identity is one of the component the other one again is maybe you know saving the your llm model on the blockchain saving your output on the blockchain that's something that's happening and at the end of the day you can somehow connect the dots with a system of checks and balances I mean it's not not perfect is not probably never be or very difficult to be really explainable AI but everything help and we can somehow understand also in history retrospect retrospectively by adding together all these components um and to that point the it's

going to be an interesting Dove tail between them so for example when you're doing kyc to open an account but you're using let's say an AI solution to perform or help the verification you're going to have to do kyc in the customer and KY AI on the AI solution so you're going to have it's they're going to dovetail and that's the whole point of you're using it to optimize but be careful and know what you're using um especially when you have sensitive uh when you're doing it for sensitive purpose like screening people or you using interacting with personal data because this touches on so many different areas of Regulation I don't think uh needs to be part of the conversation um any more questions just raise the

hand no okay then um closing remarks how do you see this uh going and who is going to be that like is the partnership with um with government institution is going to be key or actually uniting projects like you guys are doing uh is going to push this collaboration with government how I mean is it chicken and egg yeah a little bit it has to be both sides right these are heavily regulated uh heavily regulated uh uh uh Industries uh so we can talk and try to do things uh from an independent and private sector perspective But ultimately they need to be adopted right we spoke about standards for example we can have the

best standards but even in uh you know financial markets we still use GDP as the as the center of everything which is obsolete to say the least so we need we need the discussion between public public and private to be um uh to happen uh but Innovation tends to happen from the private sector so um it's definitely worth continuing in in addition to um our products we're also leveraging our balance sheet and and Venture partnering and investing in other companies um that are doing things complimentary and uh exciting in in the space so that's another thing that we're doing to like Advance the conversation and we're really interested in partnering with

other entrepreneurs and Founders and companies that could um leverage our skills collectively for for greater impact thank you m uh yeah an important remark for me is related last week I was at AI for good in Geneva conference and and I have the feeling that there's really no need to understand as well which kind of fuse cases are good for AI also realizing the fact that you know AI is very good for support for example in our case we identify green washing and it's okay to have some false positives and you know well it's not the case for every

industry where everything can be fully automatized and therefore as well again recognizing the importance of you know how can you support you know how AI can support or identifying you know together which kind of use cases is super important as as well as understanding you know the connection between blockchain and AI because I think it's some way you know to to propel the two technologies together because there's a lot of complementarity there and just Echo the thoughts there say learn lessons from one another so the ESG brought a lot of good infrastructure in terms of reporting and getting down to that granular level that we need to to demonstrate company's

performance and AI um obviously needs to follow us similar trajectory there um and avoid avoid the also the pitfalls happen in the ESG so avoid AI washing at all costs make sure that you if something takes a bit longer you need to you need to you know sacrifice or make a tradeoff for the long term do that because once uh once once AI is fully embedded and you know present across the business it's can to be very difficult to turn back so take that take the initiative early and uh it will make your life easier a long term thank you then I'll just close with my own opinion for what it's worth uh I think that we have better chances if we

unite and then like Chase the government to to support us than just going individually project by project so um you're starting with a big merger so maybe we can all learn from that you will be the most yeah thank you um thanks everyone for joining thank you