From principles to implementation – pathways forward

AI for Good Summit Explores Regulation and Governance Challenges of Generative AI

At the AI for Good summit, Tomas Lamanauskas and Gabriela Ramos discussed the burgeoning impact of generative AI and the global challenge of its regulation and governance. Lamanauskas opened the evening session by acknowledging the hard work of his colleagues and the productive outcomes of their efforts, which revolved around the global challenge of regulating and governing AI. He highlighted the establishment of a UAI office by the European Commission, marking a significant step in AI governance. The ITU, in particular, has been instrumental in creating standards across various domains, including disaster management, health, autonomous driving, and agriculture, driven by AI.

The AI for Good summit, which began seven years ago, has evolved from a solution-focused gathering to a forum that also addresses the need for safeguarding and ethical considerations in AI deployment. Ramos contributed to the conversation by praising the ITU’s role in fostering a multistakeholder approach to AI governance. She discussed the coordination between UN agencies in advancing the global digital compact and the upcoming summit of the future. Ramos also mentioned the global consensus achieved by UNESCO on the ethics of AI, which was signed by 193 countries, and the U.S. subsequently joining, bringing the total to 194 countries.

The speakers emphasised the need for responsible AI frameworks that are closely tied to ethics and human rights. They also called for interoperability among technology platforms to ensure smaller providers can participate, as well as the development of international technical standards to keep AI working for the good of humanity. There was a clear call to leverage AI to bridge digital divides and to ensure global solidarity and resource sharing.

Lamanauskas outlined ITU’s ongoing efforts, including the development of technical standards and initiatives to combat deepfakes and misinformation. He mentioned the ITU’s partnership with the United Nations University to produce a flagship AI for Good report and highlighted the AI for Good Impact Initiative, set to launch the following day, aimed at mobilising the global community to share knowledge and assist developing countries.

The summit was positioned as an inclusive and diverse platform for addressing AI challenges and opportunities. Lamanauskas thanked participants for making the inaugural AI Governance Day a success and invited everyone to engage in the discussions over the next two days of the summit. He concluded by encouraging attendees to enjoy the evening and the early Geneva sunshine, looking forward to the continued conversations that would shape the future of AI governance.

Tomas Lamanauskas:
Thank you very much, Robert, and it’s great to follow the very loyal colleagues who always finish on time, and very effectively, and there’s a great result here. So great evening, everyone, after this long day, which is now coming to a close. But I think this day actually demonstrated what a lot of us have been talking in many halls over now a year and a half, how this rise, I would say explosion, of generative AI is putting the thoughts how to regulate it, how to govern it in everyone’s minds, in everyone’s discussions, including the policy makers. And we heard a lot today. We heard a lot about the UAI, we heard about U.S. order, executive order in AI, we heard about the Seoul Safety Summit that just recently concluded, we heard about the China regulatory frameworks, we heard about G7 Hiroshima process, and actually I was very proud today when my good friend Roberto Viola from European Commission texted me that they just established actually today a UAI office as well, which they’re very proud to be the first AI office. So it’s just actually the reason why Roberto is not with us today, so I can’t actually apologize really. But also it’s interesting that that office already has a unit which is called AI for Social Girls, so we’ll probably have to claim the copyright on that one. So indeed a lot is happening, and a lot of you are calling for more things to happen. And indeed UN system, we also want to say, has been responding, has been responding not just for one year and a half, it has been responding for quite a few years. So UN system has been responding by developing standards, and we as ITU have already a long track of standards on disaster management, on health, on autonomous driving, on agriculture driven by AI, and we do that very proudly with a big part of UN system. We also help governments and countries build a capacity in artificial intelligence era. But we also provide a platform for all of us to come together. And this AI for Good summit is an example of that. And this summit again didn’t start last year. The summit started seven years ago, and it’s always useful to remember. And the summit seven years ago started as a solution summit. It started to think, to discuss, and to agree how we use AI to progress forward, to propel it. Of course, since then, we got a little bit more feared of AI. We kind of now are much more afraid of it than we were then. So now we discuss also how to really put the right guardrails, how to safeguard it. But even in that regard, we also have work that’s being doing. And of course, Gabriela and I, also very symbolic that we’re both on this stage, so because we both of I lead UN internal coordination on AI, through which recently we’ve managed the UN system, more than 40 agencies to produce the governance AI, UN system AI governance paper. And that governance paper showcases the rich, rich tapestry of all the work that UN is already doing. It demonstrates to you that there’s already more than 50 different acts and instruments and normative norms that can be applied to AI, directly applicable to areas that are close to that, from cybersecurity to data management and data flows. It also reviews the principles that have been used in the UN system, and they’re still applicable to that. For example, inclusion, and how to make sure that everyone’s included. Multistakeholder environment, where everyone needs to come together from private sector to the government’s academia. It also really looks how we can make it happen, how can we integrate things like human rights frameworks in the governance as well. So this is very rich work that we’ve already been doing, and with this I really would like to ask Gabriela to pass the word to her, Gabriela, if you could explain us more how the UN system has been responding to these challenges and opportunities in AI, and what more we can do together.

Gabriela Ramos:
Well thank you, thank you so much, and thank you all for being here. I have to first and foremost say that ITU has been playing a fantastic role, because this platform, the AI for Good, I feel brings us together, all the international institutions, but also the multistakeholder approach, which is very important. But let me tell you that Thomas and I have been leading this work, coordinating all the UN agencies on artificial intelligence, and when we started discussing going towards the global digital compact, and the summit of the future that you know is going to happen in September, we said what could be a good contribution that we can make? Well why don’t we get together to see what the whole system has done on artificial intelligence? And I have to say that we were positively impressed, because as Thomas is saying, of course, WHO, would you imagine that WHO is not using AI to see how we can improve the medical outcomes, or the health systems in general, or that UNICEF is not looking, or UNICEF or UNESCO, to the education or the children well-being, or the standards in ITU, or the FAO looking at food systems. And the fact is that we discovered this richness, and our message to you is to use that richness, because it’s there. Of course, it’s a better understanding of the interaction of AI with these very sectoral elements, but also tools and agreements by member states to see how we can advance the development and deployment of artificial intelligence for good outcomes. But we do something else that is very important, because these institutions, of course, Thomas and I were part of the secretariat, as well as Zorin and Pritam and everybody, but we work with member states, and we support member states to agree on how to handle these things. At UNESCO, we are very proud, because since 2021, we achieved a global consensus on the ethics of artificial intelligence. 193 countries signed to that consensus, and then the U.S. came back to UNESCO, so now we are 194 countries signing to the consensus, which is ethical development of AI, meaning it’s not about the technologies, it’s about how do we ensure that these technologies contribute to enhance human rights, human dignity, and positive outcomes for everybody and the well-being of people. But then the other part, which is fantastic, international institutions also help countries implement their agreements, because it’s great to have international agreements, but it’s better to implement them. And therefore, we developed a tool, the Readiness Assessment Methodology, that we are implementing in 50 countries around the world, and I recognize Chile, who used the RAM to then change the legislation on privacy, and in general, in many other areas. Morocco, I was with the Minister Massour, and she was looking into the thing. So we bring the comparative perspective of what other countries are doing to ensure that our members also benefit from this knowledge and can advance governance of AI in a much more swifter way, and that we can then compare, discuss, and provide these platforms. We also have the Global Forum on the Ethics of AI. So I feel it’s very important that we do not only think that, well, international organizations are doing interesting things, they are doing things that matter for people. They’re doing things that, of course, we hear our member states. Every time I walk and talk to a member state, the minister of a member state of any country, I just met with Indonesia, I met with Bangladesh, I met with Vietnam or Africa, we are always eager to see how we could help. Because the whole point is to learn together. These are global technologies that need global solutions, and we can do that if we bring together the partners, if we bring together the governments who are first and foremost the ones that are responsible for the good governance of AI, but also if we do it in the way that ITU has been doing it with the businesses, with the academics, because artificial intelligence is changing our lives. We need to understand better, we need to use it better, and we need to see what kind of instruments, incentives, investments, regulations we can deploy to ensure that they deliver for good. So I invite you all, take a look at, actually we produced a report that Thomas mentioned and we’re very proud of that, this is our contribution to the discussions on the Summit of the Future, and we invite you to take a look at it, it’s actually even good read. You might think now it might be very bureaucratic and boring, it’s not, it’s good read. So I invite you to take a look at that, and actually to continue this conversation because we will be learning by doing, we might not get it right the first time, but at least we have the whole world together coming to this very important discussion. So thank you so much Thomas, I’m always proud to be partnering with ITU in these endeavours.

Tomas Lamanauskas:
Indeed Gabriela, and I think I can only join you in the call of reading the paper, and I think at least it will not put you to sleep, I can promise you, so don’t just read it before the night, and now if you want to have some sleep. But also I think a few things that you mentioned really matter. First of all, it’s probably important to have both horizontal frameworks, like the ones like UNESCO’s work on ethics, but also make sure that AI is really governed and embedded in every different other areas, be it health, be it agriculture, be it disaster management as mentioned, be it transport, because those experts in those areas should harness AI for the global good and make sure that it happens. So we also, the work is still not done. When we did the recent survey, we also realised that out of 193 countries that we sent our survey to, less than a half had AI strategy. So we have this rich information, but we need to develop and deliver these results on the ground. So we really need to do that. We really need to do, though, we need to build that global governance as we talk about that, but we need to build that global governance in a really pragmatic way, because if you wait for the treaties to be signed and agreed, sometimes even if we wait for the legal acts to be issued, technology will always overtake us. So we need to be understanding how to be really pragmatic and how to make sure that our discussions, our governance frameworks, the way we act really corresponds to the speed of technology we are having, and reaffirming these discussions in these rooms, these discussions really matter. So then thank you very much, Gabriela, for really being a great partner in that. Thank you very much. Thank you so much. So indeed, now it has a time for me to summarise a little bit what’s happened today and to take stock of the discussions today and to maybe come up to some conclusions as well, and then to let you all enjoy, hopefully there’s still some sunshine in Geneva that you can enjoy early, just before the summer, and some well-deserved, you know, whoever prefers drink or something else. So some clear themes really emerged today out of this day zero of the governance day, which is actually the innovation of this year’s summit as well. So we had the experts, leaders, and other participants today discussing, and they come up with a few things that really matter. First of all, responsible frameworks matter. We need to tie AI very closely to ethics and human rights. Second, we’ve heard that people want interoperability among technology platforms, so that smaller providers and smaller players can also play in that, but also regulate your approaches. And it’s not always possible to have a one-to-one approach, but you need to have them interoperable. We also want international technical standards to keep AI working for the good of humanity, and I think we are in a good company and a good host here, and the session just before showed that we have a determination to do that. Just as importantly, they also talked, and we heard a clear call to leverage AI to bridge digital divides, but also to make sure that we don’t create new divides. That means AI divides. And finally, we heard very clearly underline the need for global solidarity and resource sharing, be it high-performance computing or other resources that can help us all achieve our own AI goals. So we need to make sure that AI leaves no one behind while we manage these risks. So these are very valuable insights, and I want to assure you that we heard you. We means to you, means broader family, means broader stakeholder community gathered here today. So first of all, from an ITU perspective, I want to mention a few points. As ITU, we are responding. We are putting technical standards at the core of our AI governance efforts. We already have 220 technical standards applicable to AI, either delivered, around 100 of them, or in development. We are also now intensifying world standards cooperation activities with ISO and IEC to set coherent international standards, and you just saw the panel just before that talked about that. We also work on combating deepfakes and misinformation, where we’re taking major steps about which you’ll hear more in this summit further. As a little bit of a curtain raise or a teaser, look out for our Friday session on Detecting Deepfakes and Generative AI on AI Watermarking to Safeguard Authenticity and Stop Misinformation. At the heart of our work lies capacity development. Our new AI for Good Impact Initiative, set for the launch tomorrow, will mobilize the global AI for Good community to share knowledge and assist developing countries. And I said we’re doing that as ITU, but we’re not doing it alone. Actually, as you heard just before, we have a big UN family that does it together. So as you also heard earlier from the ITU Secretary General, we have just released our latest report that comprehensively highlights more than 400 UN use cases on AI. I have to mention it’s 45% more than the compendium last year, you know, so that just shows that more and more is happening. Many of these are the product of collaboration among UN agencies, and this document is key to understand our collective toolkit and keep building on our successes. I’m also thrilled about ITU’s new partnership with the United Nations University, UNU, to produce a flagship AI for Good report. It will help transform the knowledge and expertise within the AI for Good platform into a valuable resource for the stakeholders. AI for Good, indeed, reflects the best of the United Nations’ renditions and innovation, and also the lively spirit of international digital Geneva. It has grown into the most inclusive, diverse, neutral, global platform available to discuss AI challenges and opportunities. And today’s AI Governance Day attendance confirms it. This morning when we kicked it off, we welcomed 45 ministers or their representatives, 25 regulators, 25 UN representatives, and over 100 representatives of industry academia to discuss in a hands-on manner around roundtables the AI governance as well. More than half of them came from developing countries. Now in CCG, we have 1,500 people in different halls discussing that and also meeting each other. This amounts to a massive, active multistakeholder community ready to tackle the challenges and opportunities of AI in today’s world. Together we have taken an important step on good AI governance. And therefore, I really would like to thank everyone for helping to make this AI governance, I have to say, inaugural AI Governance Day, a success. All of you are warmly welcomed to actively engage in contributory discussions in the AI for Good Global Summit in the next two days in these same halls and around this building. So I wish you now good night, well-deserved rest, and see you tomorrow where I think before me Robert will take over. Thank you very much and have a great evening.