The following are the outputs of the captioning taken during an IGF 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.
***
>> GUILHERME CANELA DE SOUZA GODOI: Hi everyone. It's a pleasure to be here with you all in the audience, online, but also with this fantastic group of panelists. I'm sure if you read any of these reports about the most important issues of our times, you will certainly find the issues that we are discussing here today. Some reports will say governance of Artificial Intelligence is one of the most important issues we need to deal with. Certainly, climate change, many of these reports will say is one of the most important policy issues we have in front of us. As well as is one of the most threatening risks for humanity.
Energy resources and scarce resources and water is probably another one of those issues that you will see in different reports here and there. The challenge for this gentleman and lady here is how to combine all those things together, right? And of course with you, who I hope we are going to have a very interesting conversation about that.
So I don't mean to use a cliche here but voila, Artificial Intelligence offers maybe need lets to say interesting opportunities for addressing some of these environmental risks. And we can speak about that. How the models can help us and earth and the climate scientists and the policymakers and the civil society to solve this complex equation how we solve the planetary crises we are in the middle of.
But tragically or ironically, the same technology that can help us address the issue is contributing to the problem. Because it consumes a lot of energy. And water and so on. So again, as I said it's a cliche but our job is how we can foster the opportunity and mitigate the risk.
As usual in these very complex lives we have right now, it's easy to say but not necessarily easy to do. The god news is that these people in this panel and online they do have some interesting solutions to proposal to you. And some of them are already implementing it. So since I'm optimistic by nature, when my team asked me to moderate and I saw what they prepared I was very happy that it was not only dark and terrible but it was also about how we can actually walk the talk.
So this session is a lot about this. To have a dialogue with this group of different actors here that are dedicated to think about this problem. And how we can underlying in the one hour we have with some issues. And looking at the issues with AI problem for these problem. And we will try to showcase some tools and frameworks. UNESCO with the University College of London we are going to launch very soon a very exciting issues brief called smarter, smaller, strong errors efficient generative AI in the future of digital transformation that my dear colleague Ioanna, who is online. And who couldn't be here today and also to reduce the UNESCO carbon footprint of people traveling around the world. But she's an expert on this report and you can connect with her.
But this highlights the different approaches we can take to address these issues. So I'm sure we are going to have an exciting panel. And since we need to be expedited, I will stop here and go straight to my panelists and let me start with you, Mario. Mario Nobile is the director general of the agents for a digital Italy, AGDT. And Mario let's start with what you are doing at this strategic level. How your agency is trying to cope with these not easy challenges. Bongiorno, thank you.
>> MARCO ZENNARO: Good morning and thank you for all. And our strategy rests on four pillars. Education, first of all, scientific research. Public administration. And enterprises. And these efforts aim to bridge the divide ensuring inclusive growth and empowering individuals to thrive in an AI driven economy. For the first scientific research our goal is promoting research and the advancement of AI technologies. Technologies involves really faster. So now we are dealing with agentic AI. We started with a the large language models and then the language multimodal models. And now there is a new frontier about small models. For vertical domains. So we met also with the with the enterprises organization in Italy and we are trying application about manufacturing, health, transportation.
This is important for energy consuming models. For public administration, we are trying to improve the efficiency and the effectiveness of public services. For companies, Italy is the 7th in the world export. So we must find a way to get to the applicational layer and to find concrete solutions for our enterprises. And education, first of all, I always say that now that the debate is not humans versus machines. Now the debate is about who understand, use, manage AI, versus who don't.
Okay. And in Italy, the AI strategy is we wrote it in 2024, we have also a strategic planning. The 3‑year plan for IT in public administration. And we emphasize the importance of boosting digital transition using AI in ethical and an inclusive way. This is important for us. We have 3 minutes so I go to the conclusion.
And we are stressing our universities about techniques like incremental federated learning to reduce model size and computational resource demands. This is the first goal. So this approach minimize energy consumption and we are creating the conditions to transition from brute force models, large and energy consuming, to vertical and drive foundation models with specific purposes. Health, transportation, tourism, manufacturing. This is the point.
Now I will say, technology evolves faster than the strategy. So we have a strategy. But we are dealing with agentic AI, which is another frontier.
>> GUILHERME CANELA DE SOUZA GODOI: Thank you very much, Mario. And I'm glad to hear the conclusion in terms of what you are working with your universities. Because I mean, I'm just the international bureaucrat here I don't understand anything about these things. (Laughter).
But I did read the report my team prepared and they were making recommendations to what you were saying and. One of my questions to them is this really possible? Is this really happening? So I'm glad to hear this because this is very important.
And I liked when you started talking with how you are interacting with these different stakeholder groups because we always talk about more stakeholders but we need to show how concrete it is. So thank you very much. Marco Zennaro, next Edge AI expert today the centre of theoretical physics. And he also works with UNESCO. And my children tell me why don't you do these interesting things because they don't understand what I do. And maybe not even I. But people like Marco are the ones walking the talk in UNESCO. So Marco, you have worked extensively on 10ML and energy efficient AI applications also in African contexts.
So can you tell us and especially for a person like me this is no political scientist, we don't understand anything. How you can ‑‑ how you can understand the benefits of what you are doing?
>> MARCO ZENNARO: Sure, definitely, thank you very much. And let me introduce ML it's about running machine learning models on really tiny devices really small devices and when I say small, really small. A devices with a few kilobytes of memory with really slow processor but they have two main advantages. First is they are extremely low powered. With green AI these devices consume little power. And second advantage is they are extremely low cost.
One of these chips is less than $1. And a full device is about $10. So that's of course very positive. And they allow AI or machine learning to run on the devices without the need of internet connection. So we heard during IGF a 1/3 of the world is not connected and if we want to have data from places that are remote where there's no internet connection and I want to use AI or machine learning that's a really good solution.
So you ask about application and in 2020 together with colleague from Harvard University and Columbia university we created a network of people working with tiny ML with a focus on the Global South. And now more than 6 universities in 32 different countries so we have many researchers working on tiny ML in different environment.
And they worked on very diverse application. So just to cite few, there's colleagues in Zimbabwe that worked on tiny ML for foot and mouth disease in cows. Sticking this device in the mouth of the cow and detecting the disease. And there's colleagues in Kenya working on counting the number of bees in a beehive. There’s colleagues in Peru who use tiny ML to detect anemia through the eyes in remote villages. We have colleagues in Argentina who use tiny ML on turtles to understand how they behave. So very diverse applications and any of them have impact on SDGs. And, again, using low cost and extremely low powered devices.
>> GUILHERME CANELA DE SOUZA GODOI: Thank you, this is fascinating as you said we are having in this IGF, which I think is a good thing, a lot of discussions regarding ‑‑ I mean, in this UN language how we do these things leaving no one behind. And these are very concrete examples. Because it's about the cost. It's about being low intensive on energy. So very glad to hear that. And again, to see that those things are possible. And these offers is a bit of hope because as I said in the beginning I'm always concerned that we are only looking to the side of the problems and the terrible risks but without looking to what is already happening to address the issues.
So let me move now to have a different setting to our online guests we have one speaker that is speaking from the internet space from the digital world. And he’s Adham Abouzied. And he's managing director at the Boston Consulting Group. And welcome Adham to this conversation. And I know the Boston Consulting Group is very concerned about these issues as well. You are putting a lot of effort on that.
And you yourself have worked in the intersectional of AI and climate resilience and digital innovation. And with a specific focus on the open source AI solutions. So how you can tell us in your 3 minutes the connection of these issues with the main topic of this panel that of course is the relationship with sustainability and environmental sustainability. Welcome. And over to you.
>> ADHAM ABOUZIED: Thank you very much. Honoured and very happy to join this very interesting panel and I must say I enjoyed so much the interventions of the panelists before me. I think you asked a very interesting question. I will start my answer by basically the results of a study that's been recently made by Harvard University, which is around the value of the open source wealth that exists on the ground today.
The study basically estimates that if we would recreate only one time the wealth of open source intellectual property that is available today on the internet, it would cost us $4 billion. But it does not stop there because if you assume or if you imagine that this material was not open source and that every player who would want to use it would either recreate it or pay a license, then basically you would increase this 4 billion‑dollar of cost to 8 trillion‑dollars just because of the repetition. So in a sense, having something that's open source optimizes significantly the amount of work that is required to get the value out of these AI algorithms or digital technologies in a larger sense.
So instead of doing the thing one time you will have communities that are actually contributing and building on top of the I would say intellectual creation of each other. First, I mean, to be able to get to broader impact at a much lower cost and a much lower energy cost as well. Now, from experience also and building specifically on what Mario is saying, for you to be able to implement vertically focused AI model that is a generate values across a certain system or certain value chain, you need to have some kind of system level cooperation, which mean that is a today all of the I would say AI break throughs that we have been seeing basically through very large generic foundational models it's there because they were trained on the data that's available out there on the internet, the wealth of text and images and the wealth of videos.
Basically these models are god at this generic generative tasks because of what these models have been able to see. And train on. But now for you to be able to have a meaningful impact with models that have a vertical focus and can create value across a certain system, they have to basically see and train on data and build on top of decisions that are made across a value chain. Let me give a concrete example that relates to this basically your starting speech. Let's take a energy sector today.
If you want to have or if you are seeking to have basically an AI development that rather than creating a burden on the energy system, actually it's becoming more optimized then you also need to allow AI to create value within the energy sector. Basically help optimize the decisions from the generation to the transmission to the actual usage. And if I take this energy system as an example basically currently in very few countries in the world you can see basically cooperation data exchange, IP exchange across the different steps of the value chain.
This would be essential. And if it is allowed, through policies, through protocols like when we first started with the internet with TCPIP then the vertical models can create more significant value that would allow as a big example the energy sector to optimize its income and will allow the AI algorithms not only to consume less energy by themselves but also to help the energy sector itself I would say optimize its output and become greener.
>> GUILHERME CANELA DE SOUZA GODOI: Thank you, very interesting. So at the end of the day we need to move from the lose‑lose game to the win‑win game. And you mentioned some keywords and you will talk about optimization, cooperation. So that's a very interesting segue. Let me make a remark on the open source and the open solutions that is very interesting.
We witnessed a concrete global example during the pandemic. Right? When the scientists decided to when open the vaccines were produced in record time in human history. So here we need to use the same logic to fix this huge problem. So Ioanna Ntinou is that correct? Ioanna is a post‑doctoral researcher in virtual machine learning at the Queen Mary University in London. And you work with the RAIDO project which focuses on reliable and optimized AI again the work. So again can you walk us through this specific use case from the project, where you mentioned the show this connection with efficiency and so on, over to you. And
>> IOANNA NTINOU: Yes, thank you, happy to be here. As you said, RAIDO IS reliable AI in data optimization. What we are trying to do is to be a bit more conscious when we develop AI models on the energy consumption that the models are going to use.
By the end of the project, we will have developed a platform that we are going to test it against for real life use cases we call it pilot. And but they are spanning different domains. This can be robotics, health care, smart farming and critical infrastructure. Today we will focus in one specific domain, which is energy grid. And we are workshop with two companies from Greece. One is a energy company, the national energy company called PPC and the other one is a research centre called Sir [sounds like] what we want to do is optimize a data model for forecasting of energy demand and supply.
Particularly in smart homes so they have a microgrid. What we were given was a big time series model that would predict the energy demand. Of small electronic device that we have in our house including say a lightbulb. The problem with this model is it was a bit big. It was very good in accuracy but it was quite big and it would need often retraining because as you know when we calculate the energy consumption of a device this is dependent on the decisionality. This is planned on the different habit that is a someone can have in a house.
What we did is simply acknowledge the installation. We took this model and used it a teach tore train a smaller model to produce same, more or less same results it was quite close in accuracy. But reduce the number of parameters to 60%. So we got a much smaller model. We call this field now. And that has the same performance. And we got several benefits from this process.
First we have a smaller model that consumes much smaller energy. Secondly this model is easier to deploy in small devices as we discussed before in the panel. Which is very critical because then you can tempt democratize the access to AI to a lot of houses because if you have a small model you can deploy it much easier. And also by maintaining the accuracy, we somehow have this companies have a more accurate forecast of the energy demand they will be around. And in that way, they can consume more produce energy having this into their mind.
So we are happy with that and we hope we will help the rest of the pilots get good models and have good results.
>> GUILHERME CANELA DE SOUZA GODOI: Thank you, and fascinating and it's a bit about evidence based next steps either for the companies or the policy making. I remember a few years ago before the AI times but already in the open data environment, a Minister of energy of a particular country launched the hackathon with data so people could help to improve the efficiency of energy consumption and so on in the public sector.
So transparency always sheds light sometimes in funny ways. And after they did the hackathon and the people did the run the models and so on, they found out that the ministry that consumed more energy in that country was the ministry of energy and environment. (Laughter) which was a big shame for them. But then they actually implemented some concrete policies to resolve these issues. So that's also interesting. We need to connect these conversations with the conversation of transparency and accountability and so on.
So Mark last but not least Mark Gachara, senior advisor at the Mozilla foundation. And we are going to get back to the conversation I guess of open source because Mozilla has a long experience on that. this mission of Mozilla that was always a beaconing in the conversations of the internet environment how you guys are looking into those new challenges now, in particularly the ones discussed in this panel, over to you and welcome.
>> MARK GACHARA: Thank you so much. Actually to build upon what you just shared which is very important. I would go back to the definition of trustworthy AI. What has happened more has been we think of trustworthy AI in terms of other dimensions in terms of governance. But you do not think about the transparency in terms of disclosure on the AI life cycle how much energy is being consumed.
We think about Let me let me say human rights which is very important. We should not harm people. We should think about this issue of economic impact. But you don't think about how we are harming the environment in that. And it is a bit of pick in that area. So at Mozilla one thing we've been trying to work on since 2023 is to think about the transparency. The transparency of how much energy are we consuming? So concertedly from last year, we ran a fund. A challenge where we had a number of grantees think about this particular area.
We had, for example, an organization from France called Code carbon. And within this theme of environment tall justice they were thinking when I write code how many energy am I using? Because at the end of the day force is used to develop the energy. So a developer can envelope their code. And it's open source that people can access. And they are building and using python tools and then we also have in the U.S., methane mapper. They did a hyper spectral image that is a detect emissions. We do not rely on maybe a self‑assessment or self‑disclosure, but rather we are kind of doing imaging and using AI to detect where our leaks before they harm the environment.
This is just an example of a practical use cases that are measuring. Because in management they have told us if you can't reshape you can't manage it and probably can't manage the risk. So I think being able to run this action research that can talk to policy eventually, then sheds light to make like environmental justice part of a core definition how we roll out AI solutions. Thank you.
>> GUILHERME CANELA DE SOUZA GODOI: Thank you, and super interesting and if I hear you correctly there's two important issues there may be an energy efficiency by design from the moment you are writing the code which is an interesting story also from the human rights perspective to be by design but also the capacity of having decent risk assessments for these things.
So now what is going to happen is we are going to have a second round here quickly, ping‑pong. And then we will open to you so start thinking of your questions to these fabulous people here. Mario you are also the regulator and. Regulators can use sticks and also carrots. So I guess my question for you is about the carrots.
>> MARIO NOBILE: Thank you, I will try to connect some dot. And my answer ‑‑ I have three answer for this. But I want to connect. I fully agree with the other panelists. With Adham. The first one is education. And I connect with the answer from Adham. We are writing with public consultation guidelines for AI adoption, AI procurement and AI development.
So think about public administration, we in Italy we have 23,000 public administration. Everyone must adopt. Everyone must do procurement. Some of them must develop. I think ministries like environment and energy of course. But also the international institute for welfare. And what Adham was saying before about the interaction between cloud and Edge computing. This is important. But we have also other challenges and I'm thinking about the energy model.
About which are the challenges now for a good use of Artificial Intelligence. Data quality. And breaking the silos. These two points are really important for us. So the first answer is education, collaboration, sharing. New questions about what can I do with my data? And with my data quality? What can I do about the cloud and the edge services I can develop?
The second one and the agency for digital Italy is working on it is the open innovation framework. It's about new applications. Not the classic tender about buying something. But the open innovation framework. So I can enable public administration to carry for planned procedures and facilitate the supply and the demand. In a new way. The third one is money.
So (Laughter). So the carrot is money. We are using in Italy a big amount of money from the national recovery and resilience plan. We have 69 billion Euros for ecological transition and 13 billion Euros for business digitization. Now we are working on a new framework about tax credits for enterprises for a good start of using AI in the small and medium enterprises.
>> GUILHERME CANELA DE SOUZA GODOI: Thank you very much. This is fascinating. And also finishing with the money also gives people some hope. (Laughter). But I wanted to understood line the aspect of procurement. In 99% of the UN member states the public sector is still the biggest single buyer. So if we can have decent procurement policies is already a lot. So congrats on that.
Marco, on your side of the story with your experience, if you can tell us what are the key enablers for this story? What are the drivers? If someone needs to look to that the policymakers and the scientists what do you suggest to start looking to it?
>> MARCO ZENNARO: Sure you said you want concrete suggestion. So I'll be very concrete and practical. The first is investing in capacity building in AI. Capacity building has been mentioned many times in the last few days B uh if new aspect from my side is to give priority for funding for curricula developed with global partners and not inventing a new wheel but using curricula like that we developed with the tiny network that is open and can be reuse.
The second is to promote local infrastructure for tiny ML deployment. And people need to have these devices in their hand. And often that's not easy. So subsidizing the access to the open source tiny ML tools and the hardware would be useful.
Of course this lowers the entrant barriers and stimulate local innovation with the SDG related challenges. The first is to integrate tiny ML to innovation strategies. People when they design a strategy don't forget you have this alternative model of really tiny devices running AI. So that's a component that should be included.
Next is to fund contexts of panel projects in key development sectors. We heard from Mozilla funding pilots and testing new solution well that's possible for tiny ML. And even we have seen many, many interesting applications. And funding even more would be extremely useful. And finally, facilitating regional collaboration and knowledge sharing so we had a few activities which were like tailored for a specific regions with the idea that people from the same region have the same issues. And that has been extremely successful.
Supporting the south‑to‑south collaboration and focusing on specific regions to use tiny ML to solve their common issues.
>> GUILHERME CANELA DE SOUZA GODOI: Interesting. So co‑working on many levels you said. But also interesting because participating in several discussion about the DPI and the public infrastructure and public interest infrastructures. And to be honest I have not heard a lot about what you've been saying. And I think it's an interesting way to connect the dot if this could be more and more represented a as potential solution. And Adham, let me get back to you now.
I know from where you are sitting in the Boston Consulting Group you are also looking to governance models for these issue. What are key features there of obviously in treatment. It's always a challenge but you can tell us on the governance side of the story?
>> ADHAM ABOUZIED: Yeah, very clear. I think we've been mainly looking at basically how to inspire what is the right governance for a systems level change to basically push forward , accelerate the adoption of open source data sharing intellectual and wealth sharing across different systems.
I think the most important thing, ‑‑ it's to have the I would say the right policies and the right sharing protocols across every industry. To be well designed and well enforced as you said with the carrot and the stick at the same ‑‑ at the same time. It is also very important to make sure that the different players are actually incentivized to adopt. And actually as Mario was saying earlier also have the right I would say set up the right regulatory reference for them for them to adopt on their own level and then afterwards also sharing the outcomes the insights the data with others.
We have faced so many difficulties in so many sectors while trying to apply AI applications whether it's generative or whether it's other techniques with facing the current regulation. The cloud is one. Basically having the proper data governance that classification and understanding what is sensitive, what should be on the cloud, what should not be on the cloud and under which levels or layers of security.
Even worse in several countries and emerging countries that basically do not have hyper scalers or do not have actual physical wealth of data centres implemented locally. And to have actual regulation against the data traveling their data traveling cross border wouldn't even allow that the I would say the prompts, the queries that would go up to query foundational models or other models sitting in the cloud somewhere outside of the country.
So regulating this what type of prompts should travel ‑‑ should actually cross the border? Which ones shouldn't? And for which maybe you need to go for local alternatives that are more focused and smaller models and maybe less efficient but you start there until the regulation evolves as something that is very, very important.
After all what is really, really important is to have certain rules across the players in a value chain about what that I they can share in which format and what I would say are the rights and responsibilities that go with it as it moves through the value chain. And for them to believe that actually sharing is a win‑win situation. It is not the opposite of I would say a free market competitiveness and keeping the intellectual property and conserving your competitiveness, it contributes to it. And gives you a wealth of information a layer on top of which you can develop a competitive edge.
>> GUILHERME CANELA DE SOUZA GODOI: Thank you, very interesting. And I guess Marco was with me yesterday in another panel about AI and the issue of standardization and rules appeared a lot. So we will need to deal with that rather sooner than later.
So Ioanna as you noticed the second round is a lot about lessons learned and insights for the different stakeholders. So in your case, what is what you would tell to policymakers from the perspective of what you are doing and what are key lessons to share with them?
>> IOANNA NTINOU: As you said I have more of a technical background. I think one key lesson we have worked working as a AI developer is sustainable AI is not going to emerge by default. We have to incentivize.
It needs to be actively supported. And the reason that even when I train a model, I will opt for a bigger model. I will opt for more data is the way we are actually measuring success, which is measured by accuracy. So with if we are measuring everything by accuracy and sometimes we neglect the cause that come with accuracy, we might consume more energy than what is actually needed. In some cases a small improvement comes with a huge increase in energy. And we will ‑‑ we have to be careful with this tradeoff. And the one of the first steps we need to do is to touch on simply assess how much energy is used in wildly used public models.
What I mean is we ‑‑ I will give an concrete example. There is now dpdforo [sounds like] and dips stick that might track 70 billion parameters and we don't know how much energy is used when we put a simple prompt in ChatGPT. So I think we should have some transparency in the way we report energy. And by developing standards at a prompt level it would be a great first step because it will create ‑‑ it will develop awareness and transparency. And after this we can see what can be done with this.
But without having the knowledge of how much energy is actually used, I think I think we cannot focus on legislation. super, of course it's important. The transparency generates the evidence that's need. And it's a good segue for you, Mark, the question is how can a civil society be stronger in demanding more transparency and accountability for this company.
>> MARK GACHARA: I will answer the question in two parts. I will start with the things that at Mozilla we see are missing. Right now there's a lot of money going into climate or smart agriculture or climate AI solutions.
But it's to make them more efficient, more effective as opposed to how do we mitigate the harms they are causing and Ioanna just mentioned a little bit about that. So it is important from our end to think about how the solutions we build can prevent. They can also improve transparency and then they can make also it more visible over the impact on the environment over the work that you are doing. And this could come through policy work. Through research. And also strengthening community work like strengthening community, civil society organizations that are working with communities on the ground.
I'll give on example of somebody we worked with last year. But this is on going. So in Kenya, we have the centre for justice. Governance and environmental action. Working on climate justice issues. And there's plans in Kenya to build a nuclear reactor near the sea. And one of the things that this did was to do an ecological mapping of an area in the Indian Ocean. And they used AI to do this.
They have come up with a report and they are saying that we think that this is ill advised and they have given the reasons why it's ill advised we should be focusing on using renewable energy sources because Kenya is a net producer of renewable energy.
This is what civil society can do, for example. They are pushing back. It's still a work in progress right now but you can generate evidence to be able to do advocacy. Unfortunately once the ship left the dock for civil society you are left trying to push back, which is unfortunate. But again, I come back to the three things. How do we use such kind of evidence to advance policy? How can we do action research? And how do we put money into preventative into making these issues more transparent? So that actually the taxpayers and government can actually quantify the issue in front of them. Thank you.
>> GUILHERME CANELA DE SOUZA GODOI: Thank you, very interesting. And also in the needs to do that in an unfortunate shrinking space for the civil society. And with less funds for the accountability and transparency, including for journalism. So now is the time for you. Questions, Leona also on the online space to do we have questions for the panel? We have two mics and this setting we look distant but we welcome your thoughts and questions.
And Leona from the online, don't know if you are hearing us.
>> LEONA VERDADERO: Hi, no questions yet.
>> GUILHERME CANELA DE SOUZA GODOI: While maybe you guys are getting less shy, can I ask you, Leona, to be on the screen? And give us a teaser of one minute about the report that we are going to launch soon about these issues?
>> LEONA VERDADERO: Yes, hi, good morning, everyone, can you hear me? Great and great to see the panelists here and great to be joining and to echo UNESCO's work and our work on being a laboratory of areas and trying to push the envelope and being able to define what do we mean by sustainable low resource energy efficient solutions?
Yes we've been doing this research and partnership with the University College of London where we are doing environments looking at how can we optimize the inference space of AI.
Inference is what other colleagues mentioned when we are interacting with these systems what we are uncovering is and I really love this conversation so far because most of the echoed the need to be more efficient climate conscious and looking at AI in a smaller, stronger, smarter way. Here we are looking at energy efficient technique such as optimizing the model and making them smaller with quantization and distillation and all of these technical things. But what we are seeing here is the different experiment that is a we have done actually make the models smaller and also more efficient and better performing. And what that means especially for stakeholders working in low resource settings where we have limited access to compute and infrastructure, these types of models are made more accessible to you.
So it's really also part of the answering the question what type of model, what type of AI do we need for the right type of job? Also trying to demystify the thought that bigger is better. It's not necessarily better I thought as we are seeing now in terms of there's all these large language models that are power hungry and water thirsty. Here's is where we are trying to merge two things together. One is technical research. How we are able to translate that in a policy setting. What it means for policymakers if you are thinking about exactly developing, deploying, procuring AI systems for your use cases.
You were able to by presenting evidence try to move the needle by moving eco conscious AI choices and model usages and. I think this is a concrete start to have a bigger push to look come completely on what it means to use smaller and smarter AI.
Also in the chat we will put here a sign‑uplink so you will all be notified when we launch this report. Thank you.
>> GUILHERME CANELA DE SOUZA GODOI: Thank you, Leona, for the teaser and I strongly recommend you will see the report, by the way it's also beautifully designed by one of our colleagues that is an expert in designing. So it shows tries to show also through cartoons, et cetera, in a user friendly way some of these issues that are sometimes complex to understand. Questions from here? Otherwise, I have ‑‑ please take a mic.
>> Hi, my name in Jan (?) I work in media development. I'm also a former in my earlier life I was a journalist, it was 10 years ago. And I'm used to hearing new things and I must say I'm thinking when I talk because most of you told us is new to me. And that's exciting and congratulations for the report and putting together the panel the way you have done.
And I'm trying to remember other technological revolutions we have seen in the past. And how they could be made more transparent. Like heating systems in houses. And in Germany we have passes for every house where you can see how energy efficient is how is and it's obligatory in the EU I think. And we can think of other technologies where we made it more transparent to see what it means to people take conscious decisions and that's why I'm glad again you stressed the transparency issue.
Of course next thought I have is we also tried carbon as a negative good. So we forced the industry to become more efficient by giving them certificates on their carbon emissions. And I wonder, by everything you know now so far about this topic, do you think we can take transparency on energy consumption that we can one day actually trade the right to consume energy on behalf of with AI technologies? And then by that way, forced industry to develop more small models and to be really conscious of whether as you say, we really need to take the last step that would be very expensive to get the very extreme accuracy we might not need. So not to ‑‑ in the way to advance in the direction you are point to. Maybe the person from the consulting group can answer this since they are looking at the overall strategy.
>> GUILHERME CANELA DE SOUZA GODOI: We have 5 minutes 35 seconds so it's 1 minute per person to interact with Jan's question and offer us one take away. And if you don't want to necessarily interact with January's question I will ask something related to what he said. What are the questions you think science journalists should be asking about this issue. Mario, over to you.
>> MARIO NOBILE: Well, in one minute is very tricky. But I will try. I think that the keywords are related. So we have sustainability. Calls for transparency. Calls for awareness. Calls for policies. Also government policies and I think the good question is not when but now we must think about sustainability energy consumption and the AI potential, the potential for AI to displace jobs. Everything is related.
We cannot think about energy consumption without job losses. So I think that the journalist can ask for a solution to job losses and energy consumption awareness from people. It's very complicated we need two hours and not 1 minute.
>> GUILHERME CANELA DE SOUZA GODOI: Thank you, but interesting and well that's the job of the journalists to find the time to do it. (Laughter). Mark.
>> MARK GACHARA: My question would be, what kind of models do we need in the small amount in the tiny devices you could have models that are very specific to an application. Confidence or turtles. My point is do we always need super wide models that can answer every question we have? Or is it better to focus on model that is a solve specific issues, which is useful for SDGs or for humanity in general?
>> GUILHERME CANELA DE SOUZA GODOI: Small is beautiful. Right? There was a article in the New Yorker fantastic not about this but overall small is beautiful. Adham, your one minute.
>> ADHAM ABOUZIED: I think it's very interesting. I would reiterate again on what I'm saying. I think yes. Smaller focused models is something that would create significant value. And would be much more optimized. To get there, they need to have access to data that is currently not necessarily out there.
And for this to happen, we need to have in place the policy, the governance that would make this happen because there isn't a wealth of data out there that would allow this. And honestly to be ‑‑ I mean, to consume the energy that is required for AI to deliver real value within the systems and value chains that helps develop on the SDGs and the help with the daily life is more important than consuming the small energy for the models to create videos and images on the internet.
>> GUILHERME CANELA DE SOUZA GODOI: Thank you. Ioanna.
>> IOANNA NTINOU: I think my question would be as a researcher if we focus on having smaller models if we are actually neglect all the progress that has been done so far with the large language models. All the evolution it has done so far. And then I guess the answer to this is that task‑specific small models still have a great value. And I'm talk about the actual value you see. The phones the technology there is small models. Because they are bounded from the battery life and the processor that the phone has, we can see that our phones having small task specific models meaning there is still a lot of things to learn.
Except for the innards part in terms of science and knowledge and things we can intuit as humans from pushing the boundaries of research. So I think we should also give it another type of value in this, which is actually the knowledge and learning out of this process.
>> GUILHERME CANELA DE SOUZA GODOI: Thank you. Mark.
>> MARK GACHARA: On my end to respond a bit to the question that was asked. The threat of where the most impact of the climate is in the Global South. It would be a farmer and it would be Indigenous and local communities so it would come back and say how do we have funds that support grassroots organizers and in Indigenous communities to actually build some of these solutions together. It's already been said the problems are localized.
And we can focus on local solutions. And then we need to force funding strategies that could actually look into research and creating science that actually solves these climate solutions with the local communities and. Procurement has already been mentioned, which is a good thing, public procurement. Thank you.
>> GUILHERME CANELA DE SOUZA GODOI: Thank you so much. We came to the end of this fascinating discussion, I want to personally thank you all to each one of you because I have learned a lot. I hope it was the same with the audience and those online. And thank you, Adham and Leona for your online participation and for here, Mario and Marco and Ioanna and Mark. And for all of you be listening to us attentively. And let's try to make this greener world while still trying to have a good benefits from this fantastic digital and AI revolution. Thank you so much. Enjoy the rest of the IGF.
(applause).