Planetary Limits of AI: Governance for Just Digitalisation? | IGF 2023 Open Forum #37

Martin Wimmer

The analysis explores various perspectives on digital transformation, sustainability, and the environmental impact of technology. One speaker emphasises the need for a human-centric approach to digital transformation, focusing on improving individuals’ lives and preserving the integrity of the Earth. They draw on the metaphor of the Japanese rock garden to describe our relationship with technology. Additionally, they highlight the importance of considering sustainable development goals and respecting human rights in the use of technology.

Another speaker argues that digitalisation and technology should promote sustainable development goals and uphold human rights. They point out that the German development policy supports the realisation of human rights, protection of climate and biodiversity, gender equality, fair supply chains, and other important aspects. They propose that a just transition to sustainable economies requires a nurturing approach rather than exploitative practices, drawing parallels with being a “gardener.”

However, concerns are raised about the environmental damage caused by artificial intelligence (AI). The negative sentiment towards AI’s impact on the environment is highlighted, suggesting that we are currently in a state of repair. Similarly, the negative sentiment towards the industry’s lack of concern for the environmental impact of their activities is expressed. The argument is made that industry needs to consider the environmental impact, aligning with the sustainable development goals related to responsible consumption and production.

The analysis also addresses the lag in legislation and regulation related to technology. The negative sentiment is expressed, stating that legislation and regulation are often implemented too late. The need for learning and better preparedness for future technologies is emphasised, as well as the positive sentiment towards gaining knowledge from the mistakes of the past.

The role of civil society and non-governmental organisations (NGOs) in exerting pressure is highlighted as a means to drive change. The positive sentiment towards the pressure from civil society and NGOs is expressed, suggesting that their involvement is crucial in advancing sustainability and human rights.

The transformation of the internet is discussed, with references to its evolution from interconnected networks to the oldest among digital technological artifacts. The neutral sentiment is expressed towards the internet, implying that it can neither be deemed good nor bad. Instead, the focus is on the internet’s role as a foundation for various digital technologies, with artificial intelligence being considered the most recent incarnation.

Overall, the analysis highlights the importance of considering sustainability, human rights, and the environment in digital transformation and technological advancements. It also underscores the need for a human-centric approach, better industry practices, improved legislation and regulation, preparedness for future technologies, and the involvement of civil society and NGOs in driving positive change. The varying perspectives shed light on the different aspects and challenges associated with digital transformation and its impact on society and the environment.

Audience

The analysis explores different perspectives on technology development, highlighting concerns, and advocates for a proactive approach. The concerns revolve around the necessity and impact of new technologies, with a particular focus on the harms and risks faced by certain communities. It is noted that significant investments are being made in technology development, but there is a need to address the potential negative consequences associated with these advancements.

One argument raised is the need to rethink the ideology and narrative of growth and development. There is a call to move away from the traditional approach and consider alternative ways of achieving progress. The emphasis is on the importance of responsible consumption and production, as well as considering the long-term sustainability of new technologies.

Another perspective suggests that countries from the Global South are not prioritising sustainability and climate protection over digitalisation. It is argued that these nations should focus on addressing environmental concerns and ensure that technological advancements align with sustainable development goals. This observation highlights the need for a balanced approach to technology adoption and an emphasis on considering the environmental impacts.

The analysis also highlights the existing digital divide, with the most advanced centres of research and development and influential companies predominantly located outside the Global South. This observation points to the power dynamics in the technology sector, indicating that decision-making and agenda-setting are often controlled by entities outside the Global South. This imbalance calls for efforts to bridge the digital divide and empower the Global South to have a greater say in shaping the technological landscape.

In conclusion, the analysis presents a range of perspectives on technology development. It underscores concerns regarding the impact of new technologies, calls for a re-evaluation of growth narratives, emphasises the need to prioritise sustainability, and highlights the inequality in the technology sector. The analysis also suggests that a proactive approach is necessary to address the challenges and potential negative consequences associated with technology development. Overall, it provides valuable insights into the complexities of technology’s role in society and the need for a more balanced and responsible approach.

Siriwat Chhem

This analysis examines the challenges and progress of sustainable AI in Cambodia. Cambodia has experienced impressive economic growth, with an annual GDP growth rate of 7% over the past 20 years. The country also benefits from a young population, with two-thirds under the age of 30. The availability of affordable mobile data and Wi-Fi has accelerated digitisation in Cambodia. Moreover, Cambodia has bypassed card payments and adopted mobile payments directly.

However, Cambodia currently lacks specific policies on AI and sustainable AI. The country is learning from regional models and others’ mistakes to develop its own AI framework. Civil society, represented by AVI Asian Vision Institute, plays a crucial role in Cambodia’s sustainable AI development by providing policy research and capacity building in the digital economy. The institute also focuses on Cambodia’s role as a small state in global governance.

Efficiency evaluation of AI tools and platforms is important as the misconception that AI can solve everything comes at a high cost and can create more problems. Long-term partnerships and continuous engagement are essential in addressing global issues related to AI and sustainability. However, there is a challenge of lack of follow-up and building on discussed points after high-level international conferences.

AI and sustainability are long-term journeys that require careful legislation and policy development. Backtracking or catching up from a regulatory standpoint is difficult due to the established nature of AI and sustainability. It is crucial to consider the broader implications of AI beyond just the technology itself.

In conclusion, Cambodia needs comprehensive policies on sustainable AI while capitalising on its progress in digitisation. Civil society, particularly AVI Asian Vision Institute, plays a vital role in advancing the digital economy. Evaluating the efficiency of AI tools, advocating for long-term partnerships, and focusing on sustainable solutions are crucial for sustainable AI in Cambodia.

Robert Opp

Digitalization and climate change are identified as the biggest global mega-trends. Developing countries bear a disproportionate burden of climate change and face challenges in terms of digitalization. Although digitalization presents the opportunity for positive action against climate change, it is also contributing to carbon emissions.

Environmental regulations and governance should not be sidelined in the pursuit of rapid digitalization. It is important that countries prioritize reducing data centre inefficiency and addressing the issue of e-waste. The global north, as a major contributor to technology development, has a responsibility to ensure that the environmental impact of these technologies is minimized.

Forming alliances in global digital governance is crucial. Initiatives such as the Coalition for Digital Environmental Sustainability (COBE) and the AI for the Planet Alliance aim to foster political alignment and promote sustainable approaches in the digital sphere. These alliances recognize the importance of involving diverse stakeholders including the private sector, civil society, and governments.

The value of local digital ecosystems and capacity building is emphasized for addressing sustainability issues. The global pattern of AI systems often lacks representation and diversity, and local innovators may struggle with financing, skillsets, and access to tools for building locally relevant systems. Strengthening local digital ecosystems can lead to fresh ideas and innovative solutions for sustainability.

Concerns are raised about the lack of representation and diversity in AI systems, particularly generative AI. The underlying data, or lack thereof, and the training processes contribute to this issue. It is important to address this lack of diversity to ensure that AI systems are fair, inclusive, and do not perpetuate biases or discrimination.

Developing countries may face challenges in prioritising environmental issues due to limited resources. However, it is important to recognise that the current pattern of environmental issues was created primarily by countries of the global north. It is crucial for these countries to take responsibility and work towards mitigating their impact on the environment.

Advising country partners to consider environmental implications in digitalization is a key recommendation. Technology should serve people and the planet, rather than exploiting or harming them. The process of digital inclusion and transformation should continue while not forgetting the importance of environmental considerations.

In conclusion, the extended analysis highlights the need for a balanced approach to digitalization and climate change. Environmental regulations and governance should not be overlooked, and alliances in global digital governance are crucial for promoting sustainability. The importance of local digital ecosystems, diversity in AI systems, and capacity building is emphasized. Furthermore, the responsibility for environmental issues should be acknowledged and addressed by countries of the global north. Ultimately, technology should be used as a tool to benefit both people and the planet.

Moderator – Karlina Octaviany

The IGF 2023 Open Forum 37 focused on the topic of sustainable development in relation to ICT technologies, with a particular focus on artificial intelligence (AI). The discussion aimed to address the ecological and social risks associated with the rapid digital transformation.

The panel of speakers included representatives from diverse organisations, such as the German Federal Ministry for Economic Cooperation and Development, UNDP, Mozilla Corporation, ITU, and the Asian Vision Institute. These experts shared valuable insights and examples of initiatives aiming to integrate sustainability in ICT technologies and global digital governance, specifically focusing on AI.

One important aspect highlighted during the forum was the need to limit the ecological impact of digital technologies. The panelists emphasised the growing contribution of digital transformation to greenhouse gas emissions and stressed the importance of ensuring sustainable AI development and deployment. They discussed the need for sustainable aspects to be considered in the development and deployment of digital technologies, including AI, and highlighted the role of digital transformation in addressing the planetary limits of AI.

The speakers discussed various options for action to promote the sustainable development of ICT sectors and technologies, with a specific focus on AI. They proposed measures such as the development and adoption of green ICT standards to support governments and stakeholders in developing sustainable and circular ICT systems. Examples were shared to illustrate how these standards could contribute to reducing ecological impacts and fostering sustainable practices.

Another key topic of discussion was the role of civil society and business in promoting sustainable AI. The panelists discussed the specific challenges faced by communities in Africa and Cambodia in adopting and benefiting from AI technologies sustainably. They highlighted the importance of including diverse perspectives and ensuring that the benefits of AI are accessible to all members of society.

Transparency and measurement were also highlighted as crucial factors in achieving sustainable digitalisation. The need to avoid the risk of greenwashing, where companies make false or exaggerated claims about their environmental practices, was emphasised. The discussion emphasised the importance of accurate measurement and reporting frameworks to assess the ecological impact of digital technologies and ensure genuine sustainability efforts.

The forum concluded with closing statements from each of the speakers, summarising the key points raised during the discussion. There was an overall agreement on the significance of integrating sustainability in ICT technologies and global digital governance, particularly in the context of AI. The forum provided a platform for meaningful dialogue and collaboration among stakeholders to drive positive change towards a more sustainable and inclusive digital future.

Noam Kantor

Businesses have a crucial role in sustainable AI by investing in environment-friendly partnerships. This involves seeking out and investing in or partnering with organizations that mitigate the climate emergency. Tech companies should also consider the ethical standpoint of their investments. Making products more efficient and sustainable is another important aspect of sustainability. Mozilla, for example, allows developers using Firefox developer tools to track the carbon emissions of their software. Civil society plays a significant role in educating the public about the climate impacts of technologies like AI. In Africa, sustainable technological development faces challenges such as limited funding and finance. However, initiatives like Mozilla’s Africa Emrati Project aim to address these barriers. Transparency is vital in sustainability, and digital companies should develop a transparent look at their environmental impacts. Tech regulators also have a crucial role in enforcing against deceptive greenwashing claims. Making sustainability part of product development can drive sustainable digitalization. Overall, businesses, civil society, tech regulators, and individuals all have important roles to play in promoting sustainable practices in the digital age.

Atsuko Okuda

The analysis highlights the need for greener AI and ICT development to address their negative impact on the environment. The greenhouse gas emissions generated by top telecom companies were estimated to be 260 million tons of carbon dioxide equivalent in 2021. This calls for urgent action to mitigate the environmental impact of these industries.

However, digital transformation shouldn’t be abandoned; instead, it should take environmental considerations into account. AI can play a crucial role in enhancing green transformation and weather forecasting. For example, AI can improve the predictability of demand and supply for renewable energy across a distributed grid, promoting sustainable energy practices. Additionally, AI can enhance weather forecasting, which has implications for climate action.

Another concerning issue is the significant amount of e-waste generated due to the increase in internet users. It is estimated that over 70 million tons of e-waste will be generated annually by 2023. Efficient e-waste management practices, including recycling to extract critical raw materials and promote a circular economy, are urgently needed.

Standardization and recommendations for environmental performance and e-waste management are crucial to ensure all stakeholders work towards common environmental goals.

Raising awareness among wider societal groups about the environmental impact of AI and ICT is crucial. The International Telecommunication Union (ITU) is implementing an AI project to build capacity and awareness among different stakeholders. This inclusive approach enables diverse perspectives to be considered in finding solutions to environmental challenges.

The ITU is also evaluating the environmental resilience and performance of data centers, aiming to improve their sustainability.

While AI technology offers opportunities, it should be integrated with environmental considerations to minimize negative impacts.

Addressing e-waste management requires collaboration with small and medium-sized enterprises (SMEs). An area office and innovation center in Delhi is working with SMEs and businesses in India to tackle e-waste management challenges.

Policy and regulatory mechanisms play a significant role in addressing the e-waste issue, ensuring producers take responsibility for proper e-waste management, even if they are not located in the same country as end-users.

Furthermore, proper e-waste disposal practices are essential to prevent environmental and ocean pollution.

Digital inclusion and transformation are crucial for global development. However, environmental concerns must be considered alongside these goals. Approximately 2.6 billion people are still unconnected, highlighting the digital divide. Bridging this gap while incorporating environmental considerations is essential.

To summarize, addressing the negative impact of AI and ICT on the environment requires greener development practices. Key areas of concern include greenhouse gas emissions, e-waste generation, and the digital divide. Incorporating environmental considerations into digital transformation, promoting proper e-waste management and recycling, raising awareness, and implementing policy and regulatory mechanisms are vital steps towards a sustainable future.

Moderator – Karlina Octaviany:
Hello, everyone. Thank you for joining the IGF 2023 Open Forum 37, Planetary Limits of AI, Governance for Just Digitalisation. I think all of the speakers are here. Thank you, everyone, for joining on-site and also online. We are broadcasting this event also hybrid, so I think we should gather a lot of insights and also a good balance of questions online and on-site. Thank you for coming here. My name is Karlina Oktaviani. I’m Artificial Intelligence Advisor for Fair Forward Indonesia and also Digital Transformation Centre Indonesia, Global Initiatives Dedicated to Open and Sustainable Development and Applications for Artificial Intelligence in Africa and Asian Countries. On behalf of German Federal Ministry for Economic Corporations and Development, BMZ, implemented by GIZ, I will be your moderator for today. The session will run with discussion led by the moderator and also some speakers will send also a presentation and later we have a question and answer session, so please be prepared with your questions and if you have an opinion or response, we also welcome that and we shall begin the session. The digital transformation increasingly contributes to the greenhouse gas emissions. To ensure sustainable artificial intelligence or AI, there’s a need to limit the ecological and social risk. How can we ensure that sustainable aspects are considered on the development and deployment of digital technology, such as AI, and how we can form the basis of digital transformation? In this open forum, we will discuss options for action for the sustainable development for ICT and of the ICT sectors and technologies, especially for AI. I will introduce the panelists for this session. For impulse statement, we have Martin Wimmer here, Director General Development Policy Issues, German Federal Ministry for Economic Corporation and Development, BMZ. We have on that side Noam Kether, Senior Public Policy and Government Relation Analyst of Mozilla Corporation. We have Robert Opp, Chief Digital Officer of UNDP, and joining online we have Atsuko Okuda, Regional Director, International Telecommunication Union or ITU, Regional Office for Asia and the Pacific. Also joining online, we have Chem Siruwat, Director of Center for Inclusive Digital Economy at the Asian Vision Institute and Advisor to the Council for the Development of Cambodia. To begin, please welcome impulse statement from Mr. Martin Wimmer, Director of General Development Policy Issues, German Federal Ministry for Economic Corporation and Development, BMZ. Please welcome. Give the warm welcome. We can clap your hands.

Martin Wimmer:
Thank you. Yesterday morning, I went to Ryoen-Chi. This World Heritage Site in Kyoto and yours is one of the most inspiring gardens ever built by mankind since the 15th century. It is a rock garden. Basically, it consists of 250 square meters of flat gray gravel and five islands with rocks on it. It is rectangular like a screen. The gravel representing dots. You see, it can mean anything you come up with during meditating there. And it is a metaphor for the technological design, the shaping of nature and the current five hype digital technologies, AI, quantum computing, whatever everyone is crazy about. It is a metaphor for the millions of websites in the internet and the five platforms that stand out. It is a metaphor for all the millions of users and the five founders who get all the attention and money. It encourages thinking out of the box. And whether you think the digital transformation leads to good or bad, the lesson you get from the rock garden at Ryoen-Chi is that the more you focus on the five outstanding highlights, the more you watch the rocks that steal the limelight, the more your attention will shift to the gravel. If you look long enough, if you think deep enough, it’s the gravel that makes the rocks shine. There are only 15 rocks and millions of pebbles there, but the task is to leave no one behind. For our discussion today, this could mean to emphasize the importance of a human-centric perspective. What does the big platform, the new technology, the great solution, the fascinating vision of one of our outstanding speakers mean for the people who do not stand out and do not get all of the attention at first sight? The poor, the children, women, people with disabilities, LGBTI+, people in the global south, oppressed people, indigenous people, victims of terrorism and war. You don’t need to be a sand master. It’s just common sense. Whether you are a gardener or a coder, whether you use a shovel or a server for your work, using technology, data centers, AI to change the world, nature, societies, human interaction should never be for technology’s sake, but for improving the lives of every individual living with us on this planet, and to secure the integrity of this one, our Earth, which translates into safe energy, safe water, safe resources. Don’t believe in growth. Don’t fool consumption. Don’t produce waste, which, to be clear, is the opposite of what the digital economy does most of the time, most of them. If you’re serious about carbon neutrality and a just transition of our economies to sustainable economies, we have to act as a gardener. Respect, tend, nurture, not exploit data centers. Do what the rock garden does. Remain within given boundaries. That’s why the German development policy supports the global realization of human rights, the fight against hunger and poverty, the protection of the climate and biodiversity, health and education, gender equality, fair supply chains, fair working conditions, the democratic, social, ecological, feminist, inclusive use of digitalization and technologies transfer to promote sustainable development goals worldwide. Thank you.

Moderator – Karlina Octaviany:
Give your own welcome. Please clap for Mr. Wimmer. Thank you, Mr. Wimmer. I really like the analogy of making it as an ecosystem that grows everyone’s in the AI. So, let’s begin the session. I will remind you that this is an open forum, so I will encourage people and invite people to prepare your questions, your response, your opinion. If you have any points that you want to discuss, we’re open to that. And first, we’ll go to UNDP, Robert Off, Chief Digital Officer of UNDP. So, let me ask the first question. How can we form broader efforts to integrate sustainability in ICT technologies and global digital governance, including AI? Thank you so much for having me at

Robert Opp:
the panel. Thank you, Martin, for the poetry. I knew you were going to deliver something inspirational, and you’re absolutely right about the boundaries. Couldn’t agree more. I think I would like to start with just a general reflection on the issues that we’re talking about here. Digitalization and climate, these are quite possibly the biggest mega trends that we have globally right now. They are changing everything about the world, but they’re doing it disproportionately. And so, we know that a disproportionate burden of climate change is borne by developing countries. We know that digitalization is happening at different rates globally, and developing countries are at a disadvantage when it comes to the speed of digitalization or the technology generation and things. And then, between these two concepts, there is a tremendous interaction that goes, it’s a bi-directional. So, on the one hand, digitalization presents the possibility to take dramatic, positive action against climate change. On the other hand, we know that digitalization is driving carbon emissions. It’s also contaminating soil for the extractive industries that have grown up around building chips and technology platforms, the rare earth minerals and so on. And it’s even the data center techniques of using cooling water that is not a closed-loop system, can contaminate water sources and things. So, there’s a really interesting, I mean interesting, important interaction, bi-directional interaction between the concepts. I think that one of the things from a UNDP perspective that we really work with, when we work with countries worldwide on their digital transformation, and we’re engaged, well, we have digital programs in 125 countries. We are engaged in about 50 of those countries on questions of national digital transformation. And I think that our partners are in the developing world, but they, like pretty much most countries, tend to put some of the environmental regulatory and important governance discussions on the back burner in the favor of quick digitalization. And so, one of the things that we really try to do in our approach when we work with a country, and we look at their readiness for digital transformation, or the readiness for artificial intelligence, because we do that kind of assessment as well, we try to place those questions centrally. And it’s about putting people and their rights, their right to, for economic and social rights to development, but also the rights for the environment. And we put the questions in front of countries. If you’re using data centers, are you doing that in a green way? Have you looked at optimization for efficiency? Have you looked at the carbon footprint of the digital change you’re making? Are you transparently disclosing environmental impact of technologies that you’re adopting? And like I said, our countries, our partners, are developing countries, but I think every country in the world needs to put this as a central concern, particularly those who are driving the technologies. And I think the last thing I would say is simply that if we look toward what it’s going to take for global action in this, it really is going to be that this has to become the norm, so that these are central questions. And going back to that point about disproportionate impact, I really think that we need to send the signal in the global north that is developing a lot of these technologies, that we must find ways to ensure that the environmental impact, the greenhouse gas footprint of the technologies that are being used, is seen as a priority in terms of data center efficiency, reduced e-waste, reduced contamination, and so on. So those are just a few initial thoughts around that.

Moderator – Karlina Octaviany:
Thank you so much. Give a warm welcome to Israel. And next we’re heading online. Asko Okudad, Regional Director of International Telecommunication Union, or ITU, Regional Office for Asia and the Pacific. Ms. Okudad will present a presentation on examples of green ICT standards, how can they support governments and stakeholders to develop sustainable and circular ICT, including AI. Asko, if you’re ready, you can start. Give a warm welcome to Ms.

Atsuko Okuda:
Asko. Thank you very much for giving… Thank you. First of all, I would like to thank the organizer to invite ITU to this very important meeting, and I believe that Robert also shared that the topic is very timely. We should perhaps think about our action in terms of how to ensure that the AI development, as well as ICT development, are greener. And I have a few statistics that I would like to show from the recent studies. Let me start with the GDPT and rapid rollout of AI solutions globally. And I’m sure that, you know, all the participants have been using or experimented the use of generative AI, such as chatGPT, and the power of the solutions that are in front of us. I just want to share with you, with the participants, that there are many interesting and innovative use of chatGPT. One of our ITU senior officials got married recently, and I believe that he asked chatGPT to write his marriage vow. So I hope that that was successful. But there are increasingly very interesting and widening use of chatGPT in our social life, as well as in our workplace. Now, the question, perhaps, which is very relevant to this session, is the environmental impact of an increasing use of AI, because the tool itself is not material, in a way. And it is very difficult to quantify the environmental impacts. But today, I would like to share with you the two aspects. One is the electricity consumption, and second, the green gas emission in this presentation. So as you see on the screen, I hope that you are seeing the flipping slides. Of course, the increased use of AI is supported by the increasing the transmission of data. And those data are stored, as Robert mentioned earlier, in the data center, carried by different means of telecommunication. Now, that data center, as you know, consumes lots of electricity. And as you may have heard in the other sessions, there are significant progress in making data center energy efficient. However, one study still suggests that the training of AI solutions would require 3.5 million liter of water to cool the facilities itself for the computing. And additionally, there is a study that in terms of the greenhouse gas emission, the top most telecom companies estimated to have produced 260 million tons of carbon dioxide equivalent in 2021. So there are certainly benefits, but there are some environmental impacts that we have to consider. So because of these two aspects of digital transformation and the need to green, some scholars coined the phenomena as the twin green and digital transformation, which means that the digital transformation should take into account the environmental aspects. And AI can certainly enhance many different aspects of the twin green and digital transformation. For example, AI can enhance the predictability of demand supply for renewables across a distributed grid. And of course, as you know, there are benefits to improve weather forecasting by incorporating more of the real world systems in calculations. So the question I believe is the balance that we must find between these two, the green part and the transition part, and to get best out of the twin transition. Now, coming back to the data center, there are two, as I mentioned, the components in terms of the data center operation, there is a cooling part that will be required and these are significant and largest, the energy loss in the facility and the cooling replacement for water includes the refrigerant, which could contain harmful chemicals. But in addition to that, there is globally increasing data traffic and that is generated from low and middle income countries because they are now investing more in the storage and hosting solutions to meet the increasing demand of the internet users, which are increasing in these economies. And that will require more data centers in these locations and that may consume more electricity according to the latest statistics. Now, that’s the reason why we, ITU, have been working with partners such as GIZ to ensure that the green aspects are integrated into this digital transition and digital transformation. And one entry point to ensure that is the public procurement, to make sure that in the process of procurement to establish data center or improve data centers, the green and environmental aspects are considered and taken into account. So another entry point to ensure the environmental aspects is the e-waste management and to manage the critical raw materials. As you know, there are more internet users globally, mainly in middle and low income countries, which means that there are more devices for people to connect to the internet. And by 2023, over 70 million tons of e-waste are expected to be generated annually. And as you may see on the screen, it is estimated that the storage of the expected 2025 global data sphere alone would require up to 80 kilotons of neodymium, which is about 120 times the EU 2020 demand for this material. And at the same time, the critical raw materials can be extracted from this process of recycling if we do it properly. And we hope that the member countries as well as industry, academia, and in partnership with other stakeholders, we can create a circular economy to ensure that the e-waste are discarded safely. At the same time, that will regenerate and that will recover the critical raw materials. And in addition, ITU, as you know, have been working on the standardization and recommendations to ensure that best practices are applied on these critical aspects of environments and the environmental performance. And I hope that today’s discussion will shed light on some of these topics, including the green gas emission as well as the e-waste and data center management. And finally, I just want to highlight one point that we should also perhaps encourage a wider societal groups to be aware and to be exposed to this discussion through their awareness raised on the benefits and challenges of AI solutions in their societies. And I would like to share with you as a second question to my intervention at ITU has been implementing an AI project to build capacity awareness across different stakeholder groups in these four countries supported by the government of Australia. So that was my last slide. Thank you very much. Back to you.

Moderator – Karlina Octaviany:
Thank you so much, Ashikol. Next, we go to Noam Kantor, Senior Public Policy and Government Relations Analyst of Mozilla Corporation. So first question is what’s the role of civil society and business and what are the specific challenges of communities on the Africa continent on sustainable AI?

Noam Kantor:
Thanks so much. Thanks for having me. Thanks everybody for coming. Regarding the role of business, the first thing I like to do is zoom out to consider tech companies as just companies. An example of what I mean is that one thing companies do just as companies is invest in other companies and in financial instruments. So I guess my question zero is, and really a primary question is, is a tech company investing in companies or partnering with companies that exacerbate the climate emergency? I would say that’s the bare minimum before you start thinking about the tech they’re implementing. I think another thing businesses can do is share best practices in terms of how to make products more efficient and sustainable. For example, this year we at Mozilla created a way for developers using our advanced Firefox developer tools to track the carbon emissions of the software they’re developing. So I recommend you go have a look at that if you’re interested. I think civil society can also play a really significant role in education, especially regarding the climate impacts of technologies such as AI. My Mozilla Foundation colleagues wrote a review recently of all the climate, of many of the climate impacts of the internet and AI usage, including how much energy is used when you’re on a Zoom call when the video’s on versus when the video’s off, which maybe you all have seen. And it’s been a really popular article. I think that shows that people wanna know the impacts of the tools that they’re using, but in the case of technology, that information can be really hard to find. As for specific challenges on the African continent, I have to say that I’m not on our Africa team. I do wanna tell you just a little bit about our work there, because I think the team does great work and I’m really proud of it. I just wanna echo also, first of all, the disproportionate impact of the climate emergency on the African continent, which was previously discussed. One thing we’ve done is in 2021, Mozilla partnered with AfriLabs to study the African innovation landscape. Across the continent, the study that we did with them found key innovation barriers, such as access to funding and finance, local policies to protect and enable the ecosystems, lack of access to affordable connectivity internet, which is a big one, and a general need to collaborate across the regions that they studied. Mozilla’s Africa Emrati Project is working to fight these barriers. I think many of the same barriers affect sustainable technological development in the area. But ultimately, we think that communities should be able to speak to and try to solve their own challenges with support from others. That’s why the Mozilla Technology Fund, which supports open source projects with promising approaches to solving pressing issues, recently announced that the theme for this year is AI and environmental justice. The fund will provide $300,000 to open source projects that leverage AI to make a positive impact on the environment and local communities. It includes one year of Mozilla mentorship and support, and awardees will likely be announced in early 2024.

Moderator – Karlina Octaviany:
Interesting. So if you want to, you are welcome. And if anyone wants to explore more, you can ask later about the finding. And we go online again to Tam Sriwat, Director of Center for Inclusive Digital Economy at the Asian Vision Institute and advisor to the Council for the Development of Cambodia. Tam, are you ready? Okay, he’s online. So what’s the role of civil society and business and what are the specific challenges faced by community in Cambodia and specifically on AI?

Siriwat Chhem:
Yes, thank you very much for your question. First of all, I would like to thank the organizers for inviting me on this panel of very esteemed and distinguished panelists. Just to let you know, I’ve been following the IGF for a very long time in my research. It’s always been a dream to come and attend, but unfortunately I wasn’t able to join in person, but at least I’ll be able to join the panel online. And so onto the question. Maybe I’ll start with the second part first, since we’re talking about specific challenges faced by Cambodia on sustainable AI. I think the first thing you think of when you think of Cambodia might not be related to technology or let alone sustainable AI, but maybe I can just share a little bit of the context. And so pre-COVID for the last 20 years, Cambodia had been experiencing 7% GDP growth annually, so developing extremely quickly. And I would say that within the last five years, Cambodia has gone through its own form of digital transformation. If you were to visit around five to 10 years ago, you would see that we predominantly use cash everywhere. Also making it more complicated, we are a dollarized economy, meaning that we are on dual currency, both with the USD and our local currency. And so basic things like going to the market or taking a tuk-tuk for transportation around, you would have to do these things very manually and having complications of converting currencies and so on. And basically what happened throughout the last few years, a very high digital adoption rate. We’ve been able to, let’s say leapfrog the era of using cards, credit cards, debit cards, and move straight into mobile payments, transfers and QR code payments. And so that the main reason I would say is because of Cambodia’s young population with two thirds of the population under the age of 30, a median age of 26. We have quite affordable mobile data and access to Wi-Fi within the urban population. This has allowed us to really move forward in terms of digital transformation. And so now if I can just go back to the question of the theme for today on sustainable AI, we face different types of challenges from the previous ones that were mentioned. Because we could say that we joined late to the game, our focus is really building it from the ground level up. And because we don’t have any legacy technology or any established longstanding institutions in terms of AI, we rely quite significantly on looking at the models of our regional partners, on looking at what’s being done successfully around the world, and also learning from others’ mistakes. So in terms of sustainable AI, we are, let’s say, building a strong foundation from the beginning. We don’t have any existing specific policies on AI or specific to sustainable AI. And so I think looking at regional models, what’s being done around and locally contextualizing to Cambodia’s situation is very important. And so if I could just elaborate a little bit more on the role of civil society, on behalf of our institute, on the AVI Asian Vision Institute, we are an independent think tank. And so what we’ve tried to do over the last four years in Cambodia is to provide policy research and also capacity building and training related to the digital economy. And so over the last four years, we’ve published two books, one of them on Cambodian cyberspace, another one on Cambodia’s emergent cyber diplomacy. So really giving an overview of the digital economy, what kind of role does Cambodia as a small state play in the frame of global governance? I know that will be onto the next theme and question, so I won’t talk too much about it. And so with that, I would like to close my opening remarks. Thank you.

Moderator – Karlina Octaviany:
Thank you so much. Give a warm welcome to Tim. Thank you. Thank you. Thank you. Thank you. Thank you. Okay, thank you. Okay, so I think we can move to the second round of questions. I come back to Mr. Robert Opp, what type of alliance for global digital governance are needed?

Robert Opp:
Okay. Hello, okay. No, thanks for that. I think all of these interventions so far have drawn attention to some form of angles that we’re talking about. There’s sort of private sector, there’s the civil society, the importance, there’s governments and so on, the importance of bringing together the stakeholders can’t be overemphasized. And of course, that’s what IGF is about. I think in this space, the biggest role for alliances is around alignment of purpose, alignment of intention. And I can just give a couple examples in this space of alliances that we’re involved in and that I think have some hope for the directions that we need to set globally with this. The first one is called COBE. which stands for Coalition for Digital Environmental Sustainability, and that is an initiative with the German Environment Agency, with the UN Environment Program, UNDP, the International Science Council and Future Earth. And actually recently Atsuko ITU has also joined CODES as one of the kind of core members. And CODES has engaged with over a thousand stakeholders in the last couple years that it’s existed, and it really is trying to get a few different things. One is around political alignment for the kind of these issues of the twin transitions. Then there’s sort of a set of initiatives around mitigating negative impact, and then there’s accelerating the innovations for efficiency. And so this is a kind of broad-based coalition, I would say, and there’s some action lines that are being developed now. And I think it really just highlights the importance of really coming together under common purpose. The second alliance, which is a little bit more focused on the topic at hand, is called the AI for the Planet Alliance. And that has been created by the Boston Consulting Group, UNDP and UNESCO, plus a coalition of startups called Startup Inside. And it’s a group, a kind of an odd group in a way, of players that are engaged in this issue as well, but specific to artificial intelligence. And it is really also about providing a platform where we can identify and promote innovations that are, again, driving innovations that can help us with environmental action, as well, and scaling them, as well as looking at ways to really encourage the players in the artificial intelligence space to adopt more efficient and more environmentally friendly, more sustainable approaches to their work. And these are, you know, again, things that are very multi-stakeholder in nature, open for participation of many. The organizations I mentioned are just the kind of spearhead organizers, but really open for all to be involved in. And that’s an open call for everyone who’s listening in today, as well. These can be found, I’m not going to give the websites, but they’re both, they can be googled and found online, and encourage everyone to participate. Thank you. And additional resources for our discussion

Moderator – Karlina Octaviany:
later on, you can also share. We go to Noam. So how can we move towards sustainable digitalization? Thanks. I want to talk about transparency first. I

Noam Kantor:
talked about it a little bit with the education bit before, but I do think we need a transparent look at the environmental impacts of tech tools, including AI. So, you know, sustainability reports are often a big tool towards transparency, but as we all know, there’s a spectrum of transparency when it comes to reporting. So I wanted to talk a little bit about what we do in our sustainability report. It’s maybe an example, because we hope that we’re leading the way. So my understanding is that per the greenhouse gas protocol, which is one of the reporting standards, we’re not required to calculate or report the product use emissions associated with using products like Firefox, Mozilla Hubs, and Pocket. But we want to lead by example. We want to support transparency by reporting the optional data. So we started doing it actually in 2019, and we’re hoping that it’ll encourage our peers to do the same. What we had to do, though, was we had to work with an external consultant, and we had to develop a brand new methodology, because no one had really developed a methodology for measuring the impacts of browsers, the environmental impacts of browsers. And, you know, we hope that it accounts for device emissions that can be reasonably attributed to the browser, so that, you know, it captures the work that we’re doing and what we control. So it’s possible and vital that companies report on this aspect of the work. I think the hope is that we’re showing that it’s possible, and we’re encouraging others to do so. And I think the hope is, if the impacts are too high, they should consider changing their product roadmap. Now, as I mentioned before, also related to Mozilla Developer Tools, it’s not just about the products that companies build, but also for customizable or open-source products. It’s about giving developers and users the ability to measure and reduce the emissions in the tools that they build. I also want to say that tech regulators sometimes have an interesting role to play in sustainable digitalization. A good example is the Federal Trade Commission in the United States, which is a primary, one of the primary tech regulators in the United States. But the FTC also enforces against deceptive greenwashing claims. So there’s an interesting nexus there. In fact, the FTC has just begun a once-in-a-decade update to its green guides related to deceptive environmental claims. And some commenters have specifically requested that they bolster their enforcement against certain misleading net-zero claims or sustainability claims. But there are limits to anti-greenwashing policies because they require deceptive representations in the first place. So, you know, they’re just one piece of the puzzle. But I thought that was just an interesting nexus of how different regulators can work together in the

Moderator – Karlina Octaviany:
space. Interesting. So measurement and also risk of greenwashing. Okay, we can go to the Q&A in this open forum. We welcome respectful, diverse questions and opinion. If you have any questions, please kindly raise your hand, introduce your name and organizations, and then mention your questions. Also, for participants online, I will also remind to please type your name, organization, and questions. We will select the questions to be read online. So I will give the opportunity for the people on site first. Is there any questions, opinion, or curiosity that you want to ask?

Audience:
Yeah, thank you everyone. My name is Bushree Badi. My question is around, I think much of the conversation has focused on the impacts after you start adopting these technologies or developing them, and I’m wondering like how much work is now being done to really think critically if these specific types of technologies are needed in the first place. Because it feels like we’re trying to mitigate risks that are already being, or like certain communities are being exposed to those harms and risks, and trying to kind of like put things back into the bag that shouldn’t have necessarily been implemented in the first place. And you see a lot of this type of development in places like Silicon Valley, where there’s a lot of investment that keeps going into the development of these technologies that are presented as solutions to really systematic problems that we’re facing, but fundamentally will fail to do so. And we know this as people who maybe work on this through a systematic lens or framework. So I’m wondering if you could speak to some of the work that’s being done there, because it feels like a lot of this is just responsive instead of being proactive in addressing these issues. Thank you. Interesting. We will go

Moderator – Karlina Octaviany:
around. It’s okay. Well, thank you. Well, first of all, great to be here. My name is

Audience:
José Renato, I’m from Brazil, and I have two questions actually. Maybe jumping a little bit upon her question, we started the session talking about the growth, about the possibility of thinking beyond this, let’s say, ideology, narrative, I don’t know how to put it, but of development, of growth. We use some of these terms here, so like what are the opportunities that we have to rethink this? Maybe, is there any other paradigm that we could focus on? And the second question, after, I unfortunately forgot the name of the UNDP representative. Robert, thank you very much. I apologize, I’m terrible with names. You mentioned about the role of countries from the Global South in this whole theme, and how they were sort of not prioritizing, at least as far as I could understand, the issue of like sustainability, climate protection, over the digitalization. But I would like to hear from you, and maybe if there are any other inputs would be also welcome. How is it, like, considering that we have all of this push towards digitalization, this, it is part of the whole imaginary of development, of how a development, developed economy should look like. What would be your take, considering that the most advanced search centers of research, of development, the companies that dictate most of the agenda, they’re outside of these territories. It’s like, how do you work with these countries? How do you, you could potentially work with them to some degree, either create an environment in which they can build upon, in which it’s not like, in which they’ll be, they’ll have the benefits of all of this, even when we consider that many nations who are advancing these technologies are not fulfilling these questions. So yeah, thank you so much.

Moderator – Karlina Octaviany:
Thank you so much for all the questions. So we can move to our panel, starting

Robert Opp:
with Robert. Sure, I can address particularly that that last question. In a phrase, the value of local digital ecosystems here is super important, and this is very relevant for AI. It’s relevant far beyond a sustainability question. The concern that I have, and anybody who’s spoken to me recently has heard this, because I say this over and over again, I am very concerned about the global pattern of rollout of AI systems, particularly generative AI at the moment, because I worry about the representation and diversity in technology, in the underlying data or lack of data, and in the training process as well. And I believe that one of the most important things that we can do is to look at the ways to build capacity for local digital ecosystems, so that local innovators who are, you know, innovators and entrepreneurs are everywhere, but they sometimes lack the ingredients, and you were talking about that before, Noam. They may lack the financing, they may lack the skill set or the access to skills, and they may lack the set of tools to compete globally, or not necessarily compete globally, but to actually build systems that are locally relevant, and that will actually work towards satisfying the needs of people locally, and the needs of those markets locally. And so I really think, and this will also I think benefit the sustainability agenda as well, the stronger the local digital ecosystems are in these countries around the world, the more I think we’re going to see innovative and fresh looks at how we can address the sustainability issue as well. So that would be my response to your question about, you know, the countries, and when I said countries are not necessarily prioritizing environmental issues, that’s not a criticism. That’s because developing countries have a lot on their plates right now, and need to, are desperately short of resources, and in a constrained environment where you’re trying to really think about where you’re going to put your scarce resources, it may not be the first instinct to put it into something like that, and that actually I think needs to, the light needs to be shone toward the countries of the global north, who’ve basically created this pattern, and they didn’t think about environmental concerns either. That’s why we have this issue. And so what we say is that going forward, as we work on digitalization in these countries, we advise our country partners to stay aware of the environmental considerations as part of their governance, to think about the policy and regulatory environment that needs to be there from the beginning, so that ultimately that will pay off down the road. Maybe I’ll let other panelists answer some of the

Noam Kantor:
other questions. You can go, Noam. Still on? Okay, I think probably I have the most to say on the first question, about, you know, question is when should we not implement technologies at all, given their risks and their benefits? I think it’s the golden question, and I guess I just want to talk about the ways that trust, the concepts of trustworthy AI, transparency in AI, and transparency in climate impacts, all kind of work together to create, you know, as ingredients to create, you know, hopefully responsibility here, which is that, you know, I think one of the challenges is that many, many of the products that you reference that might not be very effective relative to their risks, oftentimes people don’t know how to measure the effectiveness of those products, like that if we’re talking about an AI model, people don’t necessarily know how to talk about the robustness, the accuracy of the model, potential for bias, even though there’s been a lot of work on those things that sort of, both investors and the public and regulators are still learning, and will be learning for a long time, how to measure those things, and so I think the more that we can push on the side of trustworthy AI, the more it will be obvious to people what they’re weighing the environmental impacts against, right? If it’s obvious, you know, how trustworthy or how accurate a model is compared to what it’s claiming to do, then it’ll be more obvious, you know, is it worth it compared to the amount of energy we have to pay for, and then, you know, external effects that are impacting our

Moderator – Karlina Octaviany:
climate and economy. Thank you, we go to Martin. To your question, I would fully

Martin Wimmer:
agree, the damage is already done. AI is here, and we are only in repair mode once again, and the reason for that is the industry just doesn’t care about the environmental impact of their money-making, and legislation and regulation are way too late once again. All we can do is learn for the next technology that breaks through. We have to be better and faster, and we need the pressure from the civil society and the NGOs here. And then we go to online.

Moderator – Karlina Octaviany:
Atsuko, if you want to answer the question. Sure, thank you. I have two,

Atsuko Okuda:
maybe, examples where we can, you know, concretize and show concrete examples of how we can take into account the questions on AI benefits and challenges to the environment. One is the, you know, mainstreaming of greening questions. ITU has been working in the communication sector and digital technology for many years. And one of the questions we are increasingly receiving is to evaluate, for example, the resilience and performance of data centers. And we have conducted the assessments in a few countries in Asia and the Pacific. But in the process, we made sure that the environmental aspects and best practices applied in the process of assessment so that the recommendations include how to mitigate the negative impact on the environment. And I hope that there will be more of this integration of greening and environmental considerations in all the aspects of digital transformation and what we do. But I would also like to add the perhaps partnership that we can expand with the industry, especially small and medium-sized enterprises. And I want to give an example of e-waste management that I mentioned earlier, that increasingly there will be data that is generated through an increasing number of devices people are using. And in ITU, we have opened a new area office and innovation center in Delhi recently. And one of the topics that we are addressing with the association of SMEs and businesses in India is to encourage innovation and to make sure that the e-waste management and climate technologies are taken up and mainstreamed in the industry side so that we can make it as a successful and profitable business. And we hope that that will contribute to the circular economy. And I hope, I believe that more of these business models will be required now that AI is being rolled out very quickly. Thank you. Back to you. Thank you. We go to Cem.

Siriwat Chhem:
Yes, just for my final comment. Recently, about last week, I attended a workshop specializing in AI organized by the International Science Council. And so what we did, they invited AI experts from the Asia-Pacific region. And I would just like to share two of the outcomes from this full day discussion. And so the first point is on mindset. Currently, we have this mindset and mentality that AI should be the solution for everything. And this comes at very high costs, not only in terms of sustainability and environmental aspects, but even down to the efficiency of actually trying to solve a problem. And so what is happening is that now we’re starting to use AI to the extent that it creates more problems than it solves. And so the overall consensus from the workshop was that we should be extremely careful in evaluating and assessing how efficient AI tools and platforms and applications are being used, and whether it’s actually solving the problem more efficiently and effectively, and not in turn, creating more problems. And so the second part, which I would like to share, is on long-term partnerships. As I mentioned, we were in a room full of very, let’s say, qualified individuals from that field of expertise. And they shared that one of their challenges or the main problem is that when they convene together for high-level international conferences, or they have workshops or meetings, the time period leading up to the meeting, a lot of preparation and time is involved. All the stakeholders are engaged throughout the event. But the problem is that following up after meeting, not much is done to bring together all the important points that were discussed. So in terms of an extensive report, in terms of building long-term partnerships to build on what was discussed at those events, because addressing global issues in terms of AI and sustainability, it requires a lot of considerations. And these things cannot be solved in one day or in a one-week conference, but it really has to be taken many steps forward into the long term. So I would just like to conclude with that. Thank you. Thank you. We go to questions from

Moderator – Karlina Octaviany:
online audience. It’s Avis from Cameroon, from the Proto-JQVIS organization. One of the thorny problems in Africa remains the return of e-waste to producer. What binding mechanism can we put in place for its effectiveness? Anyone wants to answer from the panel?

Atsuko Okuda:
Ah, yeah. Atsuko, perhaps you can answer. Sure. Thank you. Thank you for this very important question. I have a question regarding the obligation of programs that Mr. Avis asked regarding the return of e-waste to producers. Of course, there are policy as well as regulatory mechanisms that could bring, but perhaps the, as I mentioned earlier in my example, perhaps that could be seen also as opportunities to work with startups as well as SMEs so that they can recycle the devices before, of course, e-waste. And perhaps that could be seen as one part of the circular economy. So I believe that, of course, returning the e-waste is one thing that could be, you know, mandated, but perhaps we can see more collaborative ways because the producers may or may not reside in your countries, right? So returning to the producer in some cases could be a challenge. So perhaps we can see it from a holistical and ecosystem point of view on what’s the best mechanism to make sure that the e-waste are not discarded in the environment and in the ocean. I’m not sure if we have sufficient time to answer this question, but I believe that this mechanism and how to do this is a very important and essential topic, I believe, for all of us.

Moderator – Karlina Octaviany:
Thank you. Back to you. Thank you, Asuko. So as your reminder, it’s already closing time for our open forum. So we’ll have a closing statement from each of the speakers. Perhaps we can go online first from Tim.

Siriwat Chhem:
Yes, thank you. So just back to our topic on AI and sustainability, I believe it is a long-term journey, as mentioned from our opening statement and all the panelists, that in certain cases they have already been established for a long period of time. And it’s difficult from the legislation and policy point of view to kind of backtrack or catch up for that matter. And so with that, I think rather than focusing too much on the technology, which is something that’s being done in the field of AI, we should focus more on the fundamentals, which are utility and also what are some of the implications. Because if we focus too much on the technology, we think that it’s a solution to everything rather than looking at the overall big picture and weighing out the pros and cons. So I would say that we should take a more big picture approach and looking into the long term rather than just focusing on solving immediately what we can do in the current state without

Moderator – Karlina Octaviany:
thinking too far ahead. Thank you. Thank you. Atsuko, you want to share closing remarks? Thank you.

Atsuko Okuda:
I want to also add a dimension on digital inclusion. As you know, according to the latest ITU estimates, 2.6 billion people are still unconnected. And I believe that this process of digital inclusion and digital transformation should continue so that those who need the digital technology and transformation can benefit from the technologies. But at the same time, I believe that we shouldn’t forget about the greening part and environmental considerations in the process. And I hope that this conversation will continue among all of us or in the expanding community globally so that we can make sure that we can mainstream the perspectives and considerations to the environment in the future in our effort to connect the unconnected and making the digital transformation sustainable. Thank you. Thank you, Atsuko. We go to Robert for closing.

Robert Opp:
I didn’t expect a closing statement, and I don’t have a closing statement, but I do have a couple thoughts. And actually, even these last couple thoughts that were offered about the digital divide and not focusing on technology, I think Cem is exactly right. The focus here should not be the technology. The focus should be on what best serves people and the planet. And I think that if we stay focused on what best serves people and the planet, you know, technology, we’re not going to stop the sort of innovation for commercialization process. But I think as we go forward in alignment around what needs to happen, we have to keep people, we have to make sure that technology is serving people, not the other way around. And it’s the same for the planet. We just, we can’t keep that cycle of the planet is here for the taking, for the purpose of technology rollout.

Noam Kantor:
It’s not about that. Thank you. I know it’s 2.31, so I’m between you all and your coffee. But yeah, this was fascinating. And I think what I’ve been able to see is efforts towards sustainable digitalization from code to cooperation on an international scale. How everyone in the policy stack, as it were, can make an impact from where they are. And it’s been great to learn about that. I hope you’ve also come away with the sense that better practices are possible in the tech space. And that, you know, there is a way to make progress on these goals, including when necessary, you know, not shipping certain products when it wouldn’t be responsible to do so. I don’t have a poem to end with, like we started with, which is sad, but probably something from Mary Oliver would be good, so you all can imagine that. Thank you. To Martin?

Martin Wimmer:
Yeah, interconnected networks, the internet, are a variable thing. They are something like Tupperware or color TV or punk rock. Ideas from the middle of the last century. People who were there at the beginning are very old now and have gray hair. The subtext of this conference, as I experience it, is to discuss what the digital transformation means to the internet. It’s old heroes, it’s old myth, old narratives, old governance structure. And while there’s still a community of people who believe in the value of the internet for internet’s sake, there might be a new generation out there who considers the internet to be just the oldest of many digital technological artifacts, AI being the most recent incarnation, which are not good or bad in itself. A matchstick firing global warming in the worst case scenario, or tools…