Let’s design the next Global Dialogue on Ai & Metaverses | IGF 2023 Town Hall #25

Raashi Saxena

After analysing the data, several key points emerge. Firstly, there are concerns surrounding misinformation in Gen-AI tools. Outdated or faulty information has the potential to harm reputations, and the emergence of doctored videos is a significant issue that can lead to gender-based violence harms. This highlights the need for careful consideration and regulation of Gen-AI tools to mitigate negative consequences.

The importance of addressing misinformation in dialogues is emphasised, as it is essential to navigate the threats and advances in AI technology. Including concerns about misinformation in dialogues fosters understanding and collaboration in finding solutions to tackle this issue.

The Indian government, in collaboration with Intel, has taken proactive steps to educate school students about AI through the initiative ‘AI for All’. This curriculum, implemented in central schools, aims to equip students with knowledge and understanding of AI concepts. Additionally, the government has partnered with startup incubators to promote conversations and podcasts about simpler AI concepts, broadening the accessibility of AI education.

Raashi Saxena, a notable figure in the field, is willing to share their AI curriculum and engage in offline discussions, demonstrating a commitment to collaborative exploration of AI.

Diverse participation in an Artificial Intelligence dialogue in India is celebrated, as it includes individuals from various age groups, including a Buddhist monk and housewives. The selection of participants from places with social turmoil and socio-political issues adds depth and perspective to the discussions, enriching the insights gained.

Information provision is highlighted as a fundamental aspect of empowerment. Concrete and accurate data enables people to make informed decisions. Facilitating access to reliable information fosters active participation and engagement.

AI discussions are seen as educational opportunities, expanding participants’ knowledge and understanding. The diverse contributors notably gain valuable insights.

The potential of AI in content moderation is acknowledged for its precision and ability to sift through large volumes of data. AI is considered a valuable tool in addressing harmful content, particularly following the increase in online presence due to the COVID pandemic and concerns about the treatment of human content moderators.

Developers, as key stakeholders in technology, should be actively included in conversations about its role in society. Their perspectives and expertise are crucial in finding solutions and addressing challenges.

Contextualising information according to local needs and languages fosters engagement and response. In India, in-person dialogues in small village settings, coupled with translation into local languages, facilitate more inclusive and fruitful dialogues.

The analysis also highlights that hate speech, misinformation, and propaganda are long-standing issues that technology has made more economical and efficient to spread. Ongoing efforts are needed to address these issues and regulate technology to mitigate their negative impact.

The inclusion of vulnerable groups, such as children and people with disabilities, is emphasised in discussions. It is important to adopt inclusive approaches that consider the needs and perspectives of all individuals, promoting a more equitable dialogue.

The significance of considering different languages in discussions is recognised, as it makes the dialogue more accessible to diverse communities and enables a broader range of voices to be heard.

Finally, the importance of adhering to dedicated time limits for discussions is emphasised to respect participants’ time and ensure efficient conversations.

In conclusion, the analysis of the data provides insights into AI, misinformation, education, and inclusivity. A balanced approach is needed to address challenges posed by technology, information provision and education are crucial, and inclusive dialogues should consider diverse perspectives. AI’s role in content moderation and the engagement of developers in conversations about technology’s impact are highlighted. Contextualisation of dialogue according to local needs and languages is essential, as are efforts to address long-standing issues. The inclusion of vulnerable groups and consideration of different languages promote a more inclusive dialogue. Adhering to time limits is also important.

Roberto Zambrana

In the analysis, it is highlighted that Roberto Zambrana has a neutral stance towards AI and expresses curiosity about what AI itself thinks. This suggests a willingness to engage in a dialogue with AI and consider its perspectives.

Furthermore, Zambrana advocates for a hybrid approach in reaching agreements on general terms and adapting certain topics, regardless of the country. This approach emphasises the importance of flexibility and adaptability in addressing various issues related to AI. This aligns with SDG 17: Partnerships for the Goals, which promotes collaboration and cooperation among different stakeholders to achieve sustainable development.

Additionally, the analysis emphasises the significance of education in understanding the concerns and frontline problems associated with AI. Zambrana recognises that involving both citizens and developers in the process can lead to better outcomes. This highlights the need for awareness, knowledge, and dialogue to ensure the responsible and beneficial use of AI.

Moreover, the analysis highlights Zambrana’s support for global dialogues as a means of overcoming barriers and achieving a balanced understanding of AI and metaverses. Such dialogues can foster collaboration and support between countries, helping them overcome challenges and realise the potential benefits of AI and metaverses. This is in line with SDG 9: Industry, Innovation and Infrastructure, which seeks to promote technological advancements, and SDG 17: Partnerships for the Goals, which emphasizes cooperation and collaboration between different stakeholders.

Overall, Zambrana’s neutral stance towards AI, advocacy for a hybrid approach, emphasis on education, and involvement of the technical community, and support for global dialogues underscore his commitment to fostering responsible and inclusive AI development. These insights serve as a reminder of the importance of considering diverse perspectives and engaging in collaborative efforts to harness the potential of AI for sustainable development.

Audience

In the analysis of the statements made by various speakers, several key points emerge. The first point of concern is the inequalities in access to and understanding of AI and ICT, which should be addressed. Certain populations face issues of inaccessibility and lack of comprehension of AI and ICT, and this disparity needs to be rectified. It is argued that there should be an effort to close the gap and ensure that everyone can benefit from these technologies on an equal level.

On a positive note, it is acknowledged that AI can be used to bring everyone along in global advancement. The potential of AI to drive economic growth and innovation is recognized, and the speakers highlight the importance of using AI to include everyone in the world’s growth. They question what can be done using AI to ensure that the benefits of global advancement are accessible to all.

Additionally, there is a need to balance the advancement of AI and Metaverse design with the prevention of potential problems. It is emphasized that while progress in AI and Metaverse is important, it should not be done at the cost of overlooking potential issues and risks. The speakers argue for a balance between moving forward and preventing problems, highlighting that hesitation to progress can hinder overall development.

The development of metaverse with helpful AI for teaching social-emotional skill lessons to international students is considered important. The speakers underline the need to design and implement a curriculum that incorporates the latest technologies, such as metaverse and AI, to provide effective education to international students. The current teaching systems are often based on standard meeting systems, and the integration of metaverse and AI can greatly enhance the learning experience and improve outcomes.

Global dialogue on AI with different stakeholders is seen as crucial. The speakers mention the importance of sharing knowledge and experiences about the internet, digital technology, and AI from various perspectives. This global dialogue can foster collaboration, learning, and the development of best practices in the field.

The positive impact of digital technology is emphasized. It is highlighted that digital technology helps people in different ways and has the potential to drive industry, innovation, and infrastructure. The speakers acknowledge the role of digital technology in advancing various Sustainable Development Goals (SDGs).

Inclusion of different experiences and perspectives in AI policy and best practices is advocated for. The speakers believe that incorporating various viewpoints and voices in the development of AI policies can lead to more inclusive and effective outcomes. It is argued that a diverse range of experiences contributes to the formulation of AI inclusion policies and the establishment of best practices.

The work of organizations like Mission Public, which engage with vulnerable sections of the population, is appreciated. The speakers commend their efforts to reach out to individuals who are not usually involved in processes such as the Internet Governance Forum (IGF). This outreach to unions and workers is seen as a positive step towards reducing inequalities and ensuring that all voices are heard.

A notable observation from the analysis is the importance of a qualitative approach to understanding the thinking behind deliberations. One speaker suggests that understanding the motivations and thoughts behind each deliberation can lead to a more comprehensive understanding of the issues at hand. It is argued that a more granular understanding of deliberations can be achieved by studying the thought process behind them, thereby fostering more effective decision-making.

Child protection and online safety emerge as critical topics in the context of AI. The speakers emphasize that AI can be leveraged to protect children and ensure online safety. However, they caution that AI can also cause harm, such as the creation of child sexual abuse material digitally. It is stressed that when discussing AI on a global and local level, child protection and online safety should be at the forefront of discussions.

A differentiated understanding of AI applications is deemed crucial. The speakers mention various applications of AI, including content moderation, combating fake news, and detecting copyright infringements. It is argued that having a nuanced understanding of these applications is essential for the effective and responsible use of AI.

The issue of AI bias potentially affecting the validity of information is raised. Concerns about bias in image recognition technologies are highlighted, illustrating how AI can perpetuate biases, particularly in gender representation. It is suggested that biases in AI models need to be acknowledged and addressed to ensure fairness and equality.

The need to strike a balance between regulation and the usage of technology is emphasized. One speaker calls for critical analysis and understanding of technology consumption, rather than relying solely on regulation or fear of using technology. The goal is to ensure responsible use of AI and technology while acknowledging the potential risks and benefits they bring.

Public participation in the implementation of artificial intelligence is seen as necessary. It is argued that involving the public and giving them a voice is crucial for the responsible development and deployment of AI technologies. A speaker highlights the importance of hearing from specialists and considers it a common responsibility to include experts from various fields in decision-making processes.

Proper governance of artificial intelligence is highlighted as essential. The speakers advocate for ensuring that AI is ethically and responsibly governed to prevent issues such as misinformation and fake news. It is emphasized that AI governance is crucial for maintaining peace, justice, and the functioning of strong institutions.

The importance of a uniformed approach in terms of digitalization and AI is highlighted. This includes the need for consistent standards and practices across different regions and countries. It is argued that a uniformed approach to digitalization and AI can help reduce inequalities and promote fair access to technology.

Overall, the speakers highlight the need to address inequalities, strike a balance between advancement and prevention, engage in global dialogue, ensure inclusivity, protect children, promote critical thinking, involve stakeholders, and govern AI and ICT properly. These points emphasize the importance of responsible and ethical development and use of AI technologies to achieve sustainable development goals and create a more equitable and inclusive society.

Antoine Vergne

The analysis explores various perspectives on artificial intelligence (AI) and its implications for society. One notable initiative is the AI for All programme, a collaboration between the Indian government and Intel, which aims to educate school students about AI. This programme is seen as a positive step towards ensuring that young people are equipped with the necessary knowledge to engage with AI technologies.

Opinions on the opportunities and threats posed by AI are divided. Around 50% of the groups believe that AI presents both opportunities and threats, while approximately 30% of the groups see it primarily as an opportunity. This reflects the complexity and multifaceted nature of AI and how it can impact different aspects of society.

There is a consensus, however, on the importance of aligning AI with human rights. Most groups agreed that prioritising human rights in the development and deployment of AI systems is essential. This reveals a shared understanding that while AI can bring immense benefits, it must be guided by ethical considerations and respect for fundamental human rights.

Another area where AI is seen as having significant potential is in research and development. The dialogues highlighted the belief that AI can generate numerous opportunities in this field. This aligns with the broader goal of SDG 9, which focuses on industry, innovation and infrastructure.

The notion of global governance of AI also emerges as a prominent theme. A significant number of participants expressed support for the idea that the governance of AI should occur at a global level. This recognition reflects the global impact of AI technologies and the need for coordinated efforts to address the challenges and benefits they bring.

Sharing experiences and knowledge about the internet, digital technology and AI from different stakeholders, countries and backgrounds is highlighted as being vital. This emphasizes the importance of diverse perspectives in shaping the development and utilisation of these technologies.

The European citizens’ panels, launched by the European Parliament, European Commission and the Council, were viewed as a crucial part of the Conference on the Future of Europe. These panels provided an opportunity for randomly selected citizens to discuss their views and wishes for the future of Europe. This inclusive approach highlights the value of citizen engagement and participation in shaping policy decisions.

Antoine Vergne stresses the need for ordinary citizens’ input in global discussions about internet governance. He highlights the importance of a more open and bottom-up approach to policymaking, allowing citizens to have an impact on policy decisions. This call for citizen involvement in governance reflects the desire for inclusivity and democratic decision-making processes.

The potential for future dialogues on AI and metaverses is also explored. The need to determine appropriate levels of governance and the importance of topic framing are discussed. Antoine Vergne supports the idea of both local and global topic framing for dialogues, recognising the value of context-specific discussions in addition to common global topics.

The analysis also highlights the significance of involving AI developers in global dialogues. By including developers in conversations about AI, not just in their professional capacity but also as citizens, a more comprehensive understanding can be achieved. This emphasizes the need to view developers and AI technology creators as both part of the solution and the challenge.

Global dialogues are seen as an opportunity to promote learning and mutual assistance among countries with different AI capacities. By sharing knowledge and experiences, countries can collectively address the challenges and maximise the benefits of AI technologies.

Inviting ambassadors from each participating country to engage in global-level reflections is considered an ideal approach. This facilitates the sharing of insights and lessons learned from national efforts and encourages international cooperation in addressing common AI-related issues.

Analysing qualitative data from citizen dialogues can present both challenges and benefits. While the process of aggregating and analysing the data may be complex, it offers valuable insights for policymakers and researchers. Artificial intelligence can play a role in making sense of the large amounts of data generated through citizen dialogues, enabling more informed decision-making.

Overall, the analysis reveals various perspectives on AI and its impact on society. It underscores the importance of education, alignment with human rights, ethical considerations, and global governance in harnessing the potential of AI. It also highlights the need for inclusivity, diverse perspectives and citizen engagement in shaping the future of AI technologies.

Antoine Vergne:
Okay. Thank you. So hello, everyone. My name is Antoine Verne. I am working at Mission Publique and we are working on citizen participation. It’s my pleasure today to be with Rashid and Roberto in that town hall to talk about citizens’ engagement and artificial intelligence and the future of it. So maybe Rashid, you want to say a word on yourself and Roberto too, and then we can give you some input and then we will have a discussion and then we can try to understand what could be the next step of such an initiative.

Raashi Saxena:
Thank you, Antoine. Hi, everyone. My name is Rashid Saxena. I’m from Bangalore, India. I was deeply involved in organizing the Global Dialogue back in 2020 on behalf of my country in India. I’m also a member of the scientific committee for the We the Internet project, which we will be discussing further. And I’m really happy to be here and I’m going to pass it on to my colleague, Roberto.

Roberto Zambrana:
Thank you very much, Rashid. I want to welcome as well to all the attendance to this session. It’s going to be very insightful. My name is Roberto Zambrana. I come from Bolivia. I was also involved in the dialogue in Bolivia a couple of years ago and happy also to be part of the scientific commission committee of We the Internet initiative. And well, with that, I think we can go.

Raashi Saxena:
Yes, moving on to you, Antoine.

Antoine Vergne:
Yes. So maybe, Roberto, you can share your screen and so we can have the kind of short presentation on looking back at what we’ve done together and then looking ahead at what we could do together. So it’s about the Global Dialogue on AI and Metaverses. Maybe you can put the next slide. So that was for the program today. And we can start with a short icebreaker. That’s always a nice way to get together. And maybe we can think about next question. That question is if you were able to ask an artificial general intelligence or something very, very advanced, one question, what would that question be? Maybe we can take 10, 15 seconds. And some of you in the room want to share with us or online what that question would be. So think about it. You are in front of a general artificial intelligence and you can ask one question, the kind of, for those who know, the Herschel’s Guide to the Galaxies, that question to the deep mind. You’re in front of deep mind and you can ask one question. What would that question be? Anyone would like to contribute with that? You want to? Sure, sure. We have one here in front.

Audience:
Okay, good morning, everybody. I’m Jane Mananiso, a member of parliament from South Africa. I’m part of the ICT group in parliament. And as well, I’m the WIP of Department of Higher Education and Training, Science, Innovation and Technology. One of my worries with regards to anything that has to do with ICT and AI, and AI, it’s the issue of a forever perpetual inequalities in terms of those who are at the peripherals and as well those who are illiterate. So I want to ask AI, what is it that we can do to make sure that we bring everybody along in terms of the advancement and transformation of a, for AI, be it AR, be it cyber security, cyber crime, and everything that has to do with human rights. What is it that we can do globally to ensure that we bring everybody along? And as we grow our countries and the world, we grow with everybody. Thank you.

Antoine Vergne:
Thank you very much. Thank you, Philippe. Very interesting question. Thanks a lot, yeah. Is there anyone else that would like to go? We have one question from Philippe, which is, can I trust you? So that would be the question that Philippe would ask, would be the trust question. Can I trust you? Other questions in the room? We have a mic there as well, if you want to go, or we can just pass the microphone. Ashley, what did you ask? Thank you.

Audience:
Thank you very much. Good afternoon. I’m Emi Tsudaka from Japan. I am working at a company called OSINT Tech. We use OSINT, we collect a lot of governmental press release in different languages into one language using AI. And we are trying to offer customers to more official and reliable information in English. Yes. And my question and my, what do you say, like question is that when you say design, let’s design the next global dialogue on AI and metaverses. And designing is very, very tricky if we try to avoid a lot of problems. But if we hesitate to move forward, we cannot move forward. So I would like to know the way you think about the balance of preventing problems, but going, move forward to have better world. So that’s my question, that’s. The balance between threats and how we need to advance. Exactly, yes. Thank you very much.

Raashi Saxena:
Thank you. That is a very valid question, given all of the Gen-AI tools that we have and the concerns around the misinformation aspects of it, whether it comes to information being outdated or having faulty information that could have reputational harms to gender-based violence harms with the advent of, I would say, revolutionizing and democratizing doctored videos. So yes, that’s a very relevant question to include in our dialogues.

Roberto Zambrana:
Yes. Any other question, maybe? Someone else, like, yes, please.

Audience:
Thank you. Hello, everybody. I am a high school teacher at a high school called Jiyugaoka Gakuen in Tokyo. And I’ve been trying to develop metaverse with helpful AI and want to teach international students take the social-emotional skill lesson. And we are now trying to develop such kind of a curriculum, but now I’m only using the standard meeting system. And I wonder how we can ask support for the AI developers or the metaverse researchers as a teacher, we need to use those latest technologies so that international students can collaborate together, but we don’t know how to ask for help to people in those research areas. So if you have a suggestion, I would like to collaborate with you. Thank you.

Antoine Vergne:
Thank you very much. I’m not sure if we have a solution, but we could ask the general AI to give us one. That would be one question to ask. Rashi, Roberto, what would you ask that?

Roberto Zambrana:
Well, I think I will ask the AI what the AI think of itself.

Antoine Vergne:
Yes, I think I would ask the same. I would ask the same.

Raashi Saxena:
From a curriculum perspective, the Indian government in their AI strategy has initiated something with, I believe, with Intel. It’s called the AI for All. So basically it’s building curriculum for school students in the central school board to give them education around that. And the Indian government also came in partnership with startup incubators to start podcasts and conversations around simpler concepts of AI. I’m happy to connect offline and maybe that’s of interest to you. But yes, there’s also, Antoine, we could also share the metaverse curriculum or rather the AI curriculum that we had with the dialogue. We’re happy to share that as well with you to keep the conversation going. Thank you very much.

Antoine Vergne:
Very good. So then maybe thanks, Rashi, for the transition. Maybe it’s time to look back at what we’ve done together in 2020. And maybe, Roberto, you can show us the next slide. Sure. So here, the idea what we would like to do is have a look back at a project we did together with many other partners. So on the left side, we have the strategic partners. On the right side, we have the strategic partners too, but in the countries. And on the left side, the strategic partners at a global level. And we formed a coalition and we did the design. So one of the questions was, how do you do the design? And we can talk about that later. We did the design and the implementation of what we called a Global Citizens Dialogue. So what it is, maybe, Roberto, the next slide. And the principle is pretty easy, is we take as many countries as possible over the world or all around the world and in each of those countries, we select a group of citizens, ordinary citizens, citizens that are non-engaged, non-expert, but are selected through random selection or through a system of snowballing to have a group which is representative of the diversity of its country. And so very important is non-expert, non-engaged. These are what we call day-to-day citizens, everyday citizens, ordinary citizens, lay citizens, whatever you want to call them. These are people that live in the country and have an experience of the internet or not, because in some of the countries, we, of course, also have people without internet connection, very important. And we gather them for one day of dialogue. And that day is normally one day all over the world. And they go through different topics. And for each of the topic, they get information. So we were talking about a curriculum. So it’s a very short curriculum in that case, but it’s about the main information on a topic, the main controversies on a topic. And then they discuss on that topic through one or two questions. And those questions guide the discussion. And then at the end, they give a collective answer to that question. So here you see the gender balance of our participants in 2020. So we had around almost 6,000 participants all over the world. And as you can see, we had a very good distribution of ages, good distribution of gender. And maybe on the next slide, you also see the distribution in terms of occupation, which more or less reflects the global population. So that was for us a check to see, okay, what we had in those rooms, in those almost 80 rooms, almost half of them virtual, because it was 2020. So it was in the height of the pandemic. But half of those dialogues were onsite in, yeah, how to say, in face-to-face meetings. So these are for the demographics. One of the session we had, one of the topic was governing artificial intelligence. And we asked the citizens to take a couple of positions and to discuss and collectively to give their opinion on the governance of artificial intelligence. One remark on that, it’s always what you will see after, it’s always what we call the collective judgment. It’s not only the individual opinion, but some of the resonance are also the resonance of the discussion of the groups. And that’s very important for us because we don’t want an opinion poll. We want to understand what people think, when they think. It’s a kind of advanced way of asking the people on complex topics. So, and one of the question we asked, and I’m sorry for the numbers below, because normally it should be 0, 10, 20, 30, but there was a glitch in the numbers. So it’s percentage. And one of the thing we asked the group was to reflect on if they thought that artificial intelligence was more a threat or an opportunity or equally. And as you say, at the end of the day, when they had discussed, the people said, okay, it’s equally an opportunity and a threat, almost half of the groups. And on top of that, around 30% of the groups also said that it was more an opportunity than a threat. So the first results we had was that generally people didn’t see AI as something very, very bad in itself. And it was rather a neutral or positive view on artificial intelligence. So next slide. Then we asked them, and that’s the big advantage of such dialogues is that you can have qualitative work. So we asked them to work on the priorities, which should be the priorities in developing AI systems and AI governance. And as you can see, it was the most, the highest one was it should be aligned with human rights. And then we asked them some questions which were closed questions. So that’s more individual questions in that sense. And here you can see also about the question of ethic and AI. And they had the feeling that, of course, there should be an ethicist involved in all that work in all of the different organizations. Maybe the next one. And that is hard to read, but just I wanted to give you the general impression of that. On that question, we asked the people to tell us, as a group, if there’s more an opportunity or threat in different fields of AI. The last one is where the people had the impression that it would bring the most opportunity and that research and development, science and research. They saw that it was one of the field where it would bring a lot of opportunities and not a lot of threat. Also to explain, all those sentences on the left that I’m sorry are not really readable, but I can give you, for example, the last one, where dilemma, and that’s why we work when we do, and because we are going to discuss about what could be a cycle, a new dialogue on that topic. It’s very important in those dialogues to phrase controversies or dilemmas, because we all know when it’s about to create public policy, to take decisions together, collective decisions, very often what we have is to solve trade-offs, to solve dilemmas, to solve trade-offs. And that’s when those deliberative processes, those citizens’ dialogues work very well. So if I take the last one, for example, or maybe, Roberto, you can start the one before. Can you, yeah, no, no, yes. So the last sentence, and AI brings advances in science and research that are not worth the huge investment needed. We should invest the money elsewhere. So that’s on the left. And on the right, AI brings a lot of breakthrough in science and research, which benefits humanity. So when you see on each of those lines, you had a controversy or a dilemma, and they had to choose. And for example, the first one, where the people thought the most harm would come, is data use is directed by those who want to get profit and exercise power, that is the left part, or data is used and organized for the common good and serves humanity. Here you see that the people all over the world had a more negative view on the use of data and how it would be done. So maybe next slide. And then we also asked, of course, the citizens to who should take decisions, governance, what should governance be done? And we are at the Internet Governance Forum. So it’s interesting to see that for AI, there was a very big part of the citizens that wanted to have a global level discussion and a global level governance for AI more than on other topic. And maybe you can show the next one. Okay, so maybe we can stop here. We can go one back. And maybe Rashi, Roberto, and I see that we also have Desiree online, want to add something on that experience in 2020.

Raashi Saxena:
Maybe we can give a chance to our colleagues online.

Roberto Zambrana:
Yes, we invite anyone online. We have Juliana, then I’m gonna answer. Yes, Juliana, yes. Okay.

Audience:
Hello, everybody. My name is Juliana from Indonesia. Happy to meet all the participants to interesting in AI and global dialogue organized by Mission Public. I think what Antoine presentation is quite clear about what the global dialogue on internet is happen. From my experience from 2020, it is a nice to hear sharing the knowledge, sharing the experience about what the internet, digital technology, and especially about AI in this conversation from different stakeholder, from different country and different background, because I know is that the detailed technology is help people in different ways. And everybody in different stakeholder, different country, different economic situation, it has different experience about the digital technology. And especially, I think what the different experience will bring to the inclusion policy and inclusion, what is it, the best practice for what we should do to have a better world. for the better application of AI in our life. I think it’s enough for me, Antoine.

Antoine Vergne:
Thanks, Juliana. So I’d say Roberto, Rashid, if you want to add something, but then after I can show a second example of a citizen’s dialogue on a related topic, and then we can open the discussion. I wanted to check if Noah is online. I’m not sure if Noah is online. Or anyone else that would like to share with us? Okay, so maybe, Roberto, you can share again the presentation. I can do the second part of the input. Sure. So in here, I will share another experience. This time, it was more done by Mission Publique and less by the coalition, but in a way, it’s a direct, it’s a child of the dialogue and of developments in Europe around citizen participation. And these were the European citizens’ panels. Maybe you can show the next slide. So the context in 21 and 22, we had in Europe a huge process, which was called the Conference on the Future of Europe. And this conference was launched by the European Parliament, the European Commission and the Council. And it was about asking citizens of Europe about their views and wishes and recommendations for the future of Europe. And this process was both at national and European level. It was both online and on-site. And one of the key pieces of that process were so-called European citizens’ panels. And those panels worked on the same principle as the dialogue with the internet, meaning we had a group of randomly selected Europeans coming from all EU countries and representing the diversity of Europe, each talking their own language. But that process was different because it was not one day, it was three weekends. So a much deeper process of discussion, but with a smaller group of people because it was 200 in each citizens’ panels. And in 22, 23, we had a new cycle of those panels with three topics, which were policies being prepared by the European Commission. The first topic was food waste because the Commission was preparing a directive on food waste. The third topic was learning mobilities. So the fact that you go abroad to learn and go back to your country because the European Commission was preparing a text on it, a program. And the second, you see it, it was about virtual worlds because the Commission was preparing a non-legislative text, an initiative on virtual worlds. Maybe you can show the next slide. So the next, so yes, basic facts, we had 150 randomly selected citizens with stratification from all over the countries in Europe, three weekends. And we had those citizens discuss with another, so it’s the photo you see on the links. And on the left side, maybe you can show the next slide, and Roberto.

Raashi Saxena:
Actually, in between that, Antoine, Desiree wanted to make a few comments. Yes. If Desiree is still online.

Audience:
I’m here. Hi. Yes, this is Desiree Milosevic-Evans. I wanted to make a few comments earlier on on the findings that you have presented in 2022. I believe that it’s very important, first of all, the work that Mission Public is doing, and that’s why we like to get engaged, to get out to the people who are otherwise not really close to the process of either national IGF or the global IGF. And this is not one of their first priorities to think about. So, from the point of concept, I really always liked how Mission Public tries to reach out to vulnerability section of population, but also of unions, of workers that are going to be really somehow affected by all the policies that we are discussing here. When you pointed out that some of them wanted regulation on a global level versus like regional level, I think it would have been also good to tease out like the motif as to how, why these deliberations happen that way. And if you could possibly somehow also quantify, you know, to understand a little bit the thinking behind it, I would personally find that useful. Of course, moderators that do it at the time of speaking to the groups really know, but I wonder in future it’s how we could present it a little bit, you know, more grain. And that was my only comment with regards to the first set of slides, but let’s continue with this.

Roberto Zambrana:
Thank you.

Antoine Vergne:
Yes, thanks, Desiree, and thanks. I thought you were online because you are also online in the room, so that’s why I wanted to give you the floor online, but you have both. You manage Ubiquiti, so congratulations, Desiree, for that. No, yes, maybe let me finish that from, and then I’ll go back to the question you had, and I think it’s a very important one for the future because indeed there is a lot of feedback on that that we had. So on the European level, the question we had and the commission asked to the citizens what was, what visions, principles, and actions should guide the development of desirable and fair virtual works? And we had them work a couple of weekends and give some recommendation to the commission. And yes, and then at the end, the output of that process was a communication from the commission about what they called Web4. So Desiree, you wanted to talk about Web3. We are with the commission, we are already at Web4 about virtual works. And this is very long, so you don’t need to read it, but it’s part of the official communication of the communication. And what is interesting is that they specifically mentioned the citizens panel as being the inspiration for their legislation. So here in terms of impact of such a citizen dialogue and the continuation of it, we can see what it can become. If you look at the last paragraph, the European commission says the citizens panel specified a set of guiding principles for desirable virtual works. And then they just list the values that the citizens had developed during the citizens panel. What I want to say with that is in 2020, we had a more bottom-up open approach trying to have impact on policymaking. The example I just showed here in 2022, 2023 was more top-down approach from a policy making body asking citizens, ordinary citizens. So you have really to imagine that the people that came to Brussels to take part to that panel were not expert, not stakeholder of internet. They had no clue about what metaverse is. They didn’t know the world. They didn’t know about virtual worlds, but they took the time, they were guided and were able to give recommendations and give the guiding principles they saw as important for the development of such a work. And with that, I would like to close the presentation saying, okay, what now? Yes, maybe the next one. So now, when you have heard that, but we first can have a discussion or we can have an open discussion in the time we have, but our motivation also to have that internal meeting with all the partners is to imagine what could be the future of such a dialogue. What could be a new version of it? Because we are still convinced and I look at Rashid, Roberto and other partners, I think we are still convinced that we need that input from ordinary citizens for the global discussion on those topics. So what could be the topic? And here, I also would like to connect back to Desiree what you were commenting is because what you said, indeed, we asked in 2020, we asked them, what should be the level of governance? So that was a framing to understand if they saw more global governance or local governance. And now, if we want to be more granular, as you said, we also should be able to understand should the topic be the same at global level or adapted to the context? I think it’s exactly what you were starting to explore, Desiree, is if we were to talk about a global dialogue on AI and metaverses, how should the topic be done and should it be a common topic for everyone or more local topic? And that’s the discussion we wanted to have with you. But I give the floor to Roberto, to Rashid to first comment on the presentation and introduce the discussion.

Roberto Zambrana:
Excellent, Antoine. We will manage to receive participants’ participation now. If you agree, now we are asking you if you can have an answer for these questions about first the topic that you think will be relevant to discuss regarding AI, artificial intelligence, and then if it needs to have a context in the country that we develop the dialogue. So those are the two questions, please.

Raashi Saxena:
I can reflect on a bit on the dialogue that we had in India. We did have a very interesting discussion and although a lot of people were not particularly subject matter expertise, I liked how we were looking at a very diverse age group. So we had participation from a Buddhist monk to housewives among the 50s. We also picked up participants from places where there is a lot of turmoil and angst. And I mean, I also come from a country which historically has the largest number of internet shutdowns. So internet connectivity is sparse or there has been internet shutdown for political reasons or otherwise. So what came out from those conversations is that people do have a lot to say and if you provide them with information that is concrete, that has the right data, we need to give them agency to be able to make their own decisions. But in short, it was a very good educational exercise for people to understand and people are, no matter what age, people are always going to be keen to participate and say what they have to say. So from that point of view, I thoroughly enjoyed the discussion as it was more of a literacy exercise which is something we all need.

Roberto Zambrana:
Great, we do have some participation now from the audience, please.

Audience:
Thank you, good afternoon everyone. My name is Katarzyna Stetiva. I represent the Polish National Research Institute and in terms of global dialogue on AI and what should be the main topics, I would definitely recommend child protection, child online safety because it’s a global, it’s a global both phenomenon and problem that there are many harms to children caused in the online environment. And digital artifacts are also stored online of what has happened to children. So AI can serve both as a tool to help in terms of, for example, finding child abusive materials in the big tail of either photos or videos and it can do a lot of harm. If you imagine digitally created material such as child sexual abuse material including visual appearance of an existing child. So we have different sides of the problem but I believe that this is the topic that should be discussed both at the global and at the local level, thank you.

Roberto Zambrana:
Thank you very much.

Raashi Saxena:
That makes a lot of sense when it comes to talking about AI being used for, you said content related to harms because it’s easier given that content moderators and the way they’ve been treated when it comes to their living wages or when it comes to looking at heinous content. Kind of, I would say contracting that to AI that could be more precise could also help in sifting through large volumes of data given that so many people have come in online after COVID. So yes, definitely AI would be a good application until there, next question.

Audience:
Yes, please. Morten, UNU-IGAV. Just to follow on that note, I agree with that but it comes back to also some of the survey results earlier to classical theme. It depends on the type of AI we’re talking about. Is it AI like chatGBT that students are using or we’re using for research or is used by students to cheat the teacher and skip learning experiences? Is it fake imagery? Is it fake news? So it comes back to this classical skill of not just having access but also the critical skills and thinking about what you consume online particularly when we’re looking at deep fakes and the like. ChatGBT, there was an interesting study done I think it was Oxford or MIT on law students. And they actually found that poor performing law students, this is bachelor level law, using tools like chatGBT actually increased their performance. Whereas top performance students dropped in performance because they were leaving things too late. So they started not being as creatively thinking about things as before. And these are university graduates. So there are these differentiations I think we need to make. And it comes back to walking into AI but also the metaverse and virtual realities with open eyes and with critical minds because there are pros and cons to these technologies. ChatGBT scrapes the internet and makes a proposal based on what are the loudest voices there. And if those loud voices are fake news or false information, well, that’s the output. And if you don’t double check as a consumer of these things then we are in a dangerous situation. But again, on the backend used cases to identify fake news, racist and discriminatory use, et cetera, et cetera. Even copyright infringements, there’s a lot of abilities there but there’s still all the classical dialogues and who are in charge of the algorithms. We’ve seen there’s been a lot of bias in image recognition and so forth. So how is that gonna come into this debate? And again, I think that starting with an educated learning and ensuring that we all critically assess what we consume and check alternative sources is part of that solution, not just regulation or fear of using the technology.

Roberto Zambrana:
Correct, yes. Please, if anyone would like to share also in online participants, please, you can raise your hands so we can allow you the mic, please. We have several participants online so you’re invited to. But we have here another comment, please.

Audience:
Okay, thank you. I think what I need to say here is to appreciate the responses after the question that I’ve asked in terms of what is it that I would want AI to respond on. And it was dealt with in depth from the issue of demographics, from the issue of public participation because at times people, they don’t think when we speak AI, you need to bring everybody along. People would speak about specifics, specialists. And I’m happy that now it is clear that when you speak about anything that speaks about transformation, you need to have a public participation. But I’m happy as well that in some of our participant, it comes out loud to say there’s a need of a continuous civic education in terms of AI. And I think what we need to do is not to shy away from the fact that AI governance, it is important. So that you deal with the issue of command and control, that nobody can just spread anything that has to do with any news, so that you deal with misinformation and fake news. So it is important that AI is governed properly well. And as well, I think that the issue of uniformity, irrespective of where you are nationally, locally, it’s something else. But having a uniformed approach in terms of digitalization and AI, it is important, not forgetting the fact that we all have different languages, but the content must not change. Because one of the things that makes us not to be in par in terms of developmental issues or studies, it is on the basis that when you come from America and I come from South Africa, our standards are not the same. And the contents change based on the issue of the country. But if we can agree that when we speak on the issue of artificial intelligence, we must remain in terms of the same content, I think we’ll deal with many issues that might affect progress in terms of 4IR revolution. Thank you.

Raashi Saxena:
Thanks, I think Dusty also had a few comments.

Roberto Zambrana:
So that will be a hybrid approach, meaning that we need to agree on general terms, independently of the country, but to adapt somehow a particular topics inside. Okay, great. Anyone else in the room, please?

Raashi Saxena:
I think Dusty had a few comments.

Roberto Zambrana:
Or online?

Antoine Vergne:
I have no question, no feedback online. And so for the moment, no.

Roberto Zambrana:
Maybe we have another comment, please. Okay.

Audience:
So I wanted to follow up on what the previous participant commented on the number of solutions, perhaps it led to that as well. With one of the questions one should be asking is what kind of AI implementation? should it be? And as it’s been mentioned earlier on, at the moment there’s like a plethora of many models of AI being developed, not just the chatGBT, but also the training data sets are being developed. And there is, you know, different kind of open source AI models, as we have heard in the main session, the father of Internet, supporting the open source, open sourcing some of these models, AI models that are being developed. But in that light, I also wanted to say that it’s important when we present these choices to people who are not expert in the fields, to also always give them some kind of understanding in trade-offs. So for example, maybe one example would be, you know, there could be an offset if only too few companies end up being the owners of the best data sets and having the more powerful algorithms. And on the other hand, open source could make many more models, but it comes down to many other things, like the size of the data set, whether it’s biased, if you do a search for who are the CEOs of hospitals in the world, it’s always a man. Is it true? No, it’s the wrong data set. And it’s not easy to fix that code in a line to say there are CEOs of hospitals in the world that are not only men, just a plastic example. But with the thought of whether it should be proprietary or open source AI models or training data sets, which are now more and more available, I think it’s also important to think about the guardrails that are built into some of these big proprietary models. That means that should be not allowed hate speech or, you know, they have these kind of constraints that are built in that are good in this. So you can control, maybe easier, regulate fewer of these instances of AI, whether they check GPT or something else. And on the other hand, open source, we think we would not be bound to just use a couple of these models. So there are these certain trade offs that in open source, it’s not open source unless you can modify. So you can, for example, modify that you allow hate speech. And then we ask ourselves, is this really artificial intelligence that is simulating human intelligence? If it’s intelligent, it should not be really suggesting propagation of hate speech and so on. And then there is a set of copyright issues as well. So there are all these, you know, questions that we could work on, because there is a rapidly growing set of developers that are making sustainable AI models and different kind of GPTs.

Roberto Zambrana:
Thank you. Thank you very much. One last round, if anyone, please. Yes, there is one comment. Thank you for a very inspiring presentation and comments. I’m

Audience:
Emi from Japan. I’m working for a private company right now, but involving educational arena as well, to provide reliable information from global governments, various governments. I really understood the participatory process is crucial in this topic. And now I have one new question that the developers should be involved in such participatory process. In order to, I feel like educating citizens as well as educating developers, maybe educating, the wording is not really what I mean, but to understand the concerns and also to understand the frontline problem is beneficial for both of us. So I feel, you know, understanding and educating and learning from both of the field is very important. And if I, so far, I don’t know such cases. So, yeah, I have my desire to know more about that. Thank you very much.

Roberto Zambrana:
Thank you very much. And indeed, a very important part of the dialogues will include the, I mean, not only the developers, but all the technical community that will be related to AI. It’s very important that. So thank you for that comment as well. I also wanted to make, in terms of people talk about… Yes. Sure. It’s from, yeah, it’s from Philippa Smith, right? Yeah. I’m just wondering whether a worldwide question might tackle digital divides and how this might impact on the understanding and use of AI from a global perspective. How can a global dialogue assist countries to support each other to overcome barriers so that there might be a balance in understanding and use of AI and metaverses? That’s a question that Philippa asks. I mean, AI is a tool to help with the dialogue, which has low connectivity. Is that the question? How can this dialogue can support to overcome barriers so that there might be a balance in understanding and use of AI and metaverses? So how the dialogue contributes to this? Do you want to take that one, Antoine?

Antoine Vergne:
Yes, thanks. But before that, I wanted to make a comment on Amy’s question about developers. And I think it’s really, really important indeed. And that’s something we need to extend the scope of that kind of dialogue, because they are part of the solution and the challenge. And maybe one example of something we managed to do in 2015, we had such a dialogue at global level on the climate agreement in Paris. So the same principle all over the world, groups of citizens and the question about the Paris agreement, information materials on that and discussed and gave their opinion. And in parallel to that, we did one process with employees from Engie. And you may know Engie is one of the big, big energy company at global level. And we had a process for their employees. So we had thousands and thousands of employees from Engie taking part in the same exact same dialogue as the citizens. And so the interesting part was they were as citizens, but also employees and stakeholders of an energy company. So it was very interesting to see that they had that double hat. I work for an energy company, but I am a citizen. And it was a very interesting thing. So thank you for reminding that indeed, having developers, having people that do also the technology is very important to address them not only as in their job, but also as citizens in such a process. So I think it’s yeah, thank you for that comment. And then we have a clarification from Philippa, I see about the. OK, and I don’t know if you see it, Roberto, but she says how countries that are more capable can assist others that might have issues through a dialogue. And so how can the dialogue assist the learning and mutual learning of different countries with different capacity?

Raashi Saxena:
We did that in India. In some places, we did two dialogues in places that are more remote where I wouldn’t say that the Internet penetration is low, but generally talking about digital literacy or these or these topics are usually not approached. And we those were the dialogues happened during the peak pandemic period. So we did have two in-person dialogues in small village settings where we trained a lot of journalists to be able to lies with and get outputs. And we realized that the format that we had might not have been the best. We might need a better way of contextualizing that information. We did localize it in terms of translating it into the local languages. India has a lot of languages, but we feel like maybe a more storytelling format with a few UI and UX experts testing out different ways to be able to evoke responses because that’s not something that they’re used to. People are not used to talking that long and pondering about those topics. So maybe there needs more time. But yes, there was something that was tried. And I’m sure there are other examples across the world that would have also worked. But also coming back to the question of developers, yes, developers need to be actually central to conversations like this to have, I would say, a more conscious and moralistic bent to this. At the end of the day, they’re humans. And also to also for all of us to reflect and see that all of these issues that we talk about, hate speech or misinformation, they’re not new phenomena that have been there because of the advent of technology. They’ve always been there. They’ve always had different modes. Sometimes you’ve had more expensive infrastructure to be able to enable these phenomena. But now you use technology which has made it more economical and cheaper and easier to, you know, spread propaganda. But yes, developers should be at the center of the conversation. Thank you for

Roberto Zambrana:
highlighting that. Yes, maybe a follow-up about a question from Felipe after she clarified about this. Yeah, we agree. She wasn’t taken as a tool. And this, we need to remember that this process is initially or mainly locally. So it’s between the citizens of each of our countries. But I will say, I don’t know what you think, Antoine, but maybe after we, as part of the results of the dialogues, we gather all the results, all the conclusions in each of the dialogues, we can have a sort of a round between coordinators of each of the countries to comment and maybe to identify which are the common topics as priorities that could be presented in the different instances that we need to share the reports, something like that. In that way, I think we can accomplish what it was suggesting and to actually have as a return this support coming from the countries that have more experience maybe in particular topics, assisting some others that don’t. So maybe that could be a good idea. What do you think, Antoine?

Antoine Vergne:
Yeah, that would be fantastic. And if I dream a bit, the next, of course, next piece would be to invite ambassadors from each country, from the participating citizens together at a global level to also reflect on their own results. And I think that needs a stronger infrastructure for the dialogue, but I think that would be fantastic to have that step and to be able to aggregate at different levels those results, because I think that’s one of the key that it’s very qualitative data. So it has the advantage of being that you can search a lot into it and understand why people say what they say. But at the same time, it’s the challenge, because indeed you have to analyze it. And maybe it’s where artificial intelligence can help indeed make sense of the data the citizens produce through such a dialogue, because until now, the analysis was human-made. And so maybe there is a full circle here to have AI help us understand what people say about AI, and that would be a nice way to have a circle and a connection between AI and citizens’ dialogues. It’s too far to the hour, so I think we can conclude if I get it right.

Raashi Saxena:
We have one question from the audience here.

Antoine Vergne:
Yes, okay. But I don’t know the timing, so I let you in the room get the last round of questions.

Roberto Zambrana:
The good thing is that we have the lunch after this, so we can have a license to extend a little

Audience:
bit. Please, Mark. Thank you, Roberto. Thank you, Antoine. Mark Carvell, Internet Governance Consultant. I was previously with the UK government. It’s not a question, really, but it’s just a point of information. And it may have cropped up earlier, because I arrived late for this session from the main plenary session on the Global Digital Compact. But I know from my association with Project Liberty, McCourt Institute, that they are participating in a focus group on metaverses at the ITU, and there are a number of working groups at the ITU on metaverses. And I think that is potentially a channel for contributing to the citizens’ aspects of this evolution of the convergence of immersive technologies with Internet technologies that’s going to be so transformative into those discussions. I think from what I understand, that they are valuable, quite wide-ranging, and Project Liberty’s particular interest is on decentralizing these technology platforms, on ensuring that they are properly respectful of ethics and rights and so on. So, I offer that as a piece of information. Yes. Conclusion here. I hope that’s helpful. Thank you.

Roberto Zambrana:
It really helps. Actually, we were talking about AI during the whole session, but of course, it wasn’t just that, but also some other emerging technologies like the metaverse. So, thank you very much, Mark, for that. I think we’re getting to the final moment of the session. And if we don’t have any other comment, maybe we can wrap up.

Raashi Saxena:
Yes, we can. Thank you. But I do believe we have one person. Maybe I’ll just take a room around and see if there’s anyone who has any last comments. Anyone at all? Anyone else wants to go? Can you give her the mic, please? Thank you.

Audience:
Thank you for giving me this opportunity to provide you with a final comment. A lot has been said on inclusion of developers and society and so on, but I do believe that it’s a strongly interdisciplinary issue. So, there must be a place for every single specialist who has something to say. And by this, I’m relating to the metaverse and instances because I originate from child protection environment. So, we have to benefit from what we know from the past and the research, and we have to check what is going on now. So, we need to make a bridge between the past, the present, and we need to listen to experts and specialists, such as, again, developers, sexologists, practitioners, policy makers. So, it’s a common responsibility, I would say. And by not including an expert of a particular field, we may simply overlook an important contribution. So, there is really, I think this room is a good example that there are many people from different environments, different angles, and we learn from each other. And this is the

Roberto Zambrana:
only way to proceed. Thank you. Correct. Correct. Yes. And of course, the learners, the teachers, and also not necessarily experts in a field, but also users of the technology.

Raashi Saxena:
Okay. There was one last comment that I also wanted to mention that we talk about children. There are other vulnerable groups, like people with disabilities, who should also be taken into account. And also, of course, different languages. And then, yes, we can go on and on. We should come to a halt. We don’t want to take away anyone’s lunchtime, but thank you so much for joining us. And yes, Roberto and I are going to be around at the IGF. I’m happy to take more questions, happy to have more discussions. And yeah, with this, we come

Roberto Zambrana:
to a close. Thank you. Yes. If you want, I don’t know, maybe Antoine would like to say goodbye as

Antoine Vergne:
well. Maybe just one. So, really, thank you for being there. But one thing is our intention is to not stop involving citizens into those discussions. So, if you’re interested in joining us in that effort of thinking about it and making it happen, we are open. We would love to discuss with you on how to do that together. I think. Excellent. Thank you. Thank you very much. Thank you. Thank you. Thank you. Thank you.