Uncategorized

by corporations, by innovators to secure that outcome. And if current trends continue, the majority of AI’s economic value risks being centered in the hands of countries and corporations in the Western economies in China. And some estimates suggest that 70 % of the value could be created and reside in those locations. And I think it’s for us in this context to think a bit about why we don’t need to accept that outcome. It’s by far means not an inevitability. And to democratize the impact of AI, it requires intentional design, it takes international collaboration, and it takes societies coming together to ensure that doesn’t happen. It also takes innovation and research, workforce development, private sector partnerships, and also trust, safety, and security.

And they’re the things we’re going to talk about on the panel today. And my colleagues are extremely well -placed. to share their thoughts and insights on those topics. So let me introduce the panel. We have Amit Chandha, Managing Director and CEO of L &T Technology Services. Good to see you, Amit.

Happy to be here.

Great to have you with us. Amanda Craig -Dekard, Senior Director, Office of Responsible AI at Microsoft. Great to have you with us. Sachin Kakar from India Site Development, Privacy, Safety and Security at Google. Good to have you with us, Sachin. Thank you for being with us. Lee Tedrick, Inaugurable AI Multidisciplinary Initiative Fellow, University of Maryland, Senior Advisor on the International AI Safety Report. Lee, good to have you with us. And last but by no means least, Julian Waits, Chief Experience Officer with Rapid7. Good to have you with us. Good to have you with us. Okay. So without further ado, let’s take a look at international and scientific research collaboration. And, Lee, let me come to you.

Let me pose. Here’s a question. Okay. And, Lee, let me pose. And the second international AI safety report was released just ahead of this conference, something that you’re very much an author of. Let’s start by hearing from you and then maybe, Sachin, I’ll bring you in. What opportunities do you see, Lee, in open international standards to address the technical challenges that we face while also building trust in AI -based systems and services? Which of these, how would you characterize those challenges and which are most critical in a developing country context?

Yeah, thanks for the question, Brad, and there’s a lot here. So the international AI safety report that I worked on with a panel of about 100 experts was just released. And one of the key takeaways from the report is that while we have made a lot of progress over the past year in evaluations and developing evidence, there’s still a long way to go. There’s a gap. And I think, you know, internationally. International standards organizations and similar efforts is a good way to work together to try to fill some of the gaps. ISO has already released one standard, 42 ,001, which is a good start, but we need to accelerate this, and we need to also recognize the fact that standards and evaluation metrics, you know, there’s a tension.

On the one hand, we want them to be able to apply across borders because we want to enable companies to have responsible technology flow across borders. But on the other hand, it’s really important because we all differ in terms of language and culture that we need to be able to customize them for different cultures, norms, languages. And I think, you know, the standards organizations will continue to play an important role. I spent a year working at NIST, the U .S. National Institute of Standards and Technology. One of the NIST projects is working on what we call the zero draft of trying to create a draft that we could then feed into the ISO process, and NIST is trying to collect stakeholder input into that draft.

And I think, you know, more globally, you know, efforts like the Hiroshima AI process, there are sort of all these pre -standards efforts where different stakeholders across different regions can work together. And I think that the ACs, the AI safety institutes across different countries and how they can coordinate. So I think there’s a lot of work to be done, but I think there’s a lot of avenues where we can collaborate together and make sure that we’re addressing the needs of everybody around the globe. Thank you.

Yeah, thanks, Steve. Very well covered. If I can add just a few more points. I think one of the challenges we see is copy pasting the regulations or standards from, you know, international markets to local markets may not always work. So localizing them, understanding the needs and constraints of the local area. Google launch Indiq GenBench. It’s a test bench for fine tuning. And assessing the. LLM models for local languages, supporting 29 Indian languages, 12 scripts, and 4 language families. So that shows an example of how we need to localize things. The second point is one -time audit or certification may not work as AI evolves. We need a continuous scanning and auditing to make sure we avoid any temporal drift in these standards and the applications.

So Sachin, let’s build on that. How do governments and developers collaborate in a way that we get the outcome that everyone desires, which is not to see the developed markets race ahead of developing countries? What does that collaboration need to look like?

Yeah, that’s an interesting question. I think at highest level, the way we think to bridge the gap between AI divide is to move away from traditional, traditional transfer approach. to more co -creation where developers and government coming together and and the underlying goal is that standards and regulations are seen as enablers and equalizers not as barriers or compliance hurdle so three specific dimensions in which we believe developers and government can collaborate and Google specifically focuses on number one is open source frameworks and interoperability and standards second capacity building and third is workforce upskilling and research I’ll quickly unpack each one of them so starting with open source frameworks AI is not new to Google we have been working on AI for past decades and remember Alpha fold and we were the first one to share the transformer paper on which all the LLMs are built when we were building AI we were also focusing on best AI practices and safety practices on AI And we have open sourced all the best practices to keep AI safe.

Safe SAIF, secure AI framework is something we have shared outside. And it is important to understand supply chain risk. And India’s digital transformation is characterized by DPI, the digital public infrastructure on which Aadhaar and UPI are built. So they can actually leverage some of these secure AI framework to make sure the malware attacks and the vulnerability in open source components are taken care. Now, standards is one thing. The collaboration goes beyond to adoption of them. And Google has co -built the COSI, Coalition of Secure AI Framework, with various industry partners. And this is what we are expanding in APAC, including India. Now, we are also committed to capacity building. With the government. And which means we need to provide tools and infrastructure, not just standards.

So we are proactively sharing the threat intelligence. We are building tools like SynthID and sharing with the community abroad. SynthID is a watermark technique which goes into the text, image, video, audio, and it can tell you whether it is AI -generated content. So some of these tools are also helping us to make sure our commitment towards standards goes into actual adoption. And finally, upskilling workforce, digital literacy, working with government to make sure the vulnerable section of the society, like elderly and teenagers, are aware of some of these challenges. And giving grants to institutes like IITs to push the frontier of research, like PQC, post -quantum cryptography, are other areas of collaboration between AI developers and the government and academia.

Let me just ask you both a question. Is there a trade -off between setting global standards and regulation? ensuring the right environment for innovation and collaboration?

Oh, yeah, that’s right. And that’s where you can start with the global regulations but then adapt them to the local constraints. Like we have bandwidth constraints in India. We have linguistic diversity. And therefore, the global standard should not become a hurdle for the young startups in India. Rather, they become co -creators in enabling the innovation that can happen and then evolve from there. So it’s a creative tension, and I think the best way is to be adaptive in this situation and eventually evolve to the international standard.

How do you see this interplay, Lee?

Yeah, I think, I mean, kind of in my work, you know, both in government, academia, and I spent 30 years working with the private sector, I think sort of figuring out the standards and the standards that are in place and the values that are in place. evaluation techniques is really key. You know, how are we going to evaluate these systems so we can, they can meet a certain threshold of safety. And then I think the question kind of comes in, you know, afterwards, once we know what it is, you know, should there be regulation or not? You know, I worry a lot of times that when we go too quickly toward the regulation, you know, the best of intentions may be there, but, you know, the technology is moving so quickly, regulators don’t necessarily know how to style the regulations to achieve the goal.

And I think sort of working from the bottom up with the science, developing the evaluation technique, taking into account that we do need to socialize, you know, customize for local markets is really important. And then we can get to the question of, well, should there be a regulation or not? And that’s where, you know, different countries may have different answers, but at least we’re working from a common technical framework and evaluation framework to assess systems. Thank you.

Thank you both. Let’s make a shift to… The conversation towards more public -private… collaboration, which I think we know is at the heart of driving the success that everybody’s looking for. And Sachin was talking a little bit about capacity building. Maybe we focus on those two elements. And Amanda, I’ll come to you and then to Amit. So there’s a persistent skills gap in AI. It’s very apparent and a lot’s being done to try and bridge that here in this country. How are your, has your organization, and I’ll come to you Amit with the same question, how are your organizations grappling with that challenge and also collaborating with government to help to narrow that skills gap?

Thank you. Yes, skills gap is really important. We see it as part of the sort of foundational infrastructure for what we need to work on together as Microsoft with other industry partners, government partners, other local partners. It’s going to take a whole community really working together to do this at scale. And just to take a step back for a moment briefly before I talk more specifically about skills, you know, we kind of see this as part of a holistic effort where you kind of need to support all of the enabling infrastructure for AI deployment, kind of from from the infrastructure layer all the way through sort of realizing value in local use cases. So we actually published on Wednesday a blog from our president, Brad Smith, our chief responsible AI officer, Natasha Crampton, where we talk about sort of five areas where we’re really focused on investing to kind of close the gaps between AI diffusion and the global north and global south.

So we talk about, like, hard infrastructure investment, right, in terms of connectivity, AI compute capacity, scaling is the second part of that plan. And the third part is really thinking about multilingual, multicultural AI capability. And the fourth is really working with local partners on local AI deployment and really what we can learn and what’s going to serve local communities, also what we can learn through that process around how we need to adapt the technology so it’s ready for those local use cases. And then really measuring diffusion so that we actually understand how things are going and how we can do that. And then really measuring diffusion so that we actually understand how things are going and how we can do that.

And then really measuring diffusion so that we actually have really informed interventions. And then really measuring diffusion so that we actually have really informed interventions. And then really measuring diffusion so that we actually have really informed interventions. So that’s the kind of holistic approach that we’re thinking about for public -private partnership. And looking at skilling more specifically, we actually have a new sort of initiative that we launched last July at Microsoft called Microsoft Elevate, which is really bringing together a number of ways that we engage with a community that is going to also be part of skilling everyone at scale, so sort of nonprofit communities, schools, and actually ensuring that they’re equipped with the technology itself, so with cloud compute access and with access to AI.

And then we are coupling that with investments in skilling. So we have made some big -number commitments around how we are really trying to do this at scale. I would say specifically for India is, you know, we last year, early last year, we made this commitment to scale up 10 million Indians by 2030. This year, we upscaled 5 .6 million Indians, and so we actually doubled that commitment to 20 million people by the end of 2030. And one of the ways that we’re doing that is we’re actually, we just announced this week a new Elevate for Educators in India program where we’re partnering with local schools, with vocational institutes, with higher education institutions to sort of teach the teachers, right? So you can actually work at scale, and we’re working with a number of Indian government ministries in this program to figure out, you know, what, how we can ensure that we have tailored programs for all of those different communities and that we’re thinking holistically about how.

You know, we, across those different sort of educational paths, are really meeting people where they are and equipping them to kind of do the next powerful thing with AI.

Thanks, Amanda. And as a business, L &T Tech Services, I mean, part of L &T originating here in India, but now very much involved in global markets. How are you tackling this in terms of addressing the skills gap?

Sure. So thank you. So before I go to skill gap, I do want to make a point on the regulation part. I do believe that too much of regulation can stifle innovation as well. So we’ve got to be careful on how much we do and where do we take it. And then the second part, of course, is to do regulation of traffic control in Delhi for our next event that we have. I think all of us will agree. Let’s get down to skills in a second now. I had to say that because it was a mess in the last two days. I’ve got pictures of myself in an auto rickshaw as well. So if we get down to skill gap, I want to address this three ways.

So I am responsible. I run a company which is potentially India’s first, engineering intelligence company with about 25 ,000 employees. I’ve been CEO for five years. When I took over, we used to be about 15 ,000 employees. We’re about 14 now, we’re about 25 ,000 employees. So, we look at skill gap and I look at skill levels. Three things you have to think about. Whatever work we’re doing in engineering consulting today, I want to say 40 to 50 % of that is new and built in the last five years, did not exist. I also want to say that whatever we are doing today, 60 % will be gone in about three to five years time. That’s the rate and pace of change. So, while my colleague from Microsoft spoke about skilling school stem as well as colleges, we’re doing two different things to stay current with the changing dynamics or three things.

One, we are actually reaching out to colleges. In the last year of their curriculum and we are making sure that the curriculum is going to be in the last five years. in India is contextual to what the industry needs. So we are sending our employees to teach. We are using CSR hours. We are doing all of that to build that up. We are actually participating with NASSCOM as well to be able to do that in the skill development. The second thing we are doing is upskilling our own employees. Now, again, in a developed economy, it’s very simple that you hear these layoffs that happen all the time and they are not because people don’t have work but because the skill is redundant.

So let’s go ahead and get a new set of skills. In an Indian context, my colleague here spoke about that very nicely. You can’t cut and paste. You fire a thousand people, you will actually end up spending half your working hours plus more with the labor commissioner here locally. You can’t do that. So you have to be able to skill people up while they are in the workforce. Now, one thing is developing curriculum, developing modules for them. to go through but the second part is actually making them do it so and normally in a consulting company you would send people to get get coached and do upskilling when they are not billable we actually doing it while they are billable because when they become non billable that’s not when you want it you want it before that right so that’s and it’s a major shift in how we’ve been operating the third thing that we are tracking as an engineering and a technology company is how much of personal time is the employee spending on technology development efforts beyond billing hours to the client so you come in and spend 40 hours right and that’s what you normally work now if you spend another three hours to write a technology paper you file a patent you actually go speak at a symposium all that is towards technology effort beyond billable hours.

The percentage of workforce within the company five years ago that did that was at 19%. Today, that number stands at 52 % of our workforce spends time, personal time to go spend time on technology beyond billable hours. And the net result of that has been we used to file 100 patents per year. We have gone from there, sorry, we used to file 50 patents per year. We have gone to filing 200 patents per year. So the point is that so again, summarizing, one, reach out to the local ecosystem and do it and spend the last year with them. That’s the hook in. Second, upskill the workforce within. And third, beyond just money, find a bigger purpose like technology or betterment of human race with technology to motivate your workforce to actually spend time on that.

And I think that’s what we’ve been doing and we think will be helpful. One last thing and we keep discussing India. But if I look at the US today, and I’ve lived there for 27 years now, is we will need schools to start mandating a certain level of STEM education that has to be done. Today, both my boys went to public schools in Virginia. I can tell you that in some schools, it’s broken. And we don’t do that in the US. We don’t do it in parts of Europe. We will continue to look at different countries for skills. And that is not where we want to be in 20 years time. I’m sorry. Jump in. Jump in, Julius.

I was going to agree with what you just said. Because Rapid7, like your company, of course, we’re a software company. We’ve basically mandated the use of agentic technologies by our employees, especially the ones in developing countries or countries that aren’t as developed as the United States. What I would tell you also on the education system, which is unique to the US, which is what makes India special. And that’s why we’re in such a wonderful place. It’s because of the technology. we’re so far behind, we’re forced to use labor in other societies that appreciate the use of STEM technology and where it’s embedded in the way that they learn. We have no choice. If we didn’t have foreign workers in the U .S., we would fall behind the rest of the world.

You don’t hear that too often.

Let me just probe a little bit on this. How much is carrot and how much is stick when you’re looking to upskill the workforce and bring them into more of an AI mindset? You’ve got a very bold program at Microsoft reaching across colleges, but you’re also active, I know, in creating the capabilities within the workplace. How much of this, to both of you, is carrot or stick? I was at a dinner in D .C. a few weeks ago where the head of a large media group had told his team they had to be two times more productive by the end of 25 using AI. to stay in their roles and 10 times more productive by the end of 26.

That was an expectation. But it was set very much as a minimum standard and goal. They were putting training programs in place, but there was a clear metric to achieve. What’s your perspective based on how you’ve seen this work?

You mean internally?

Either within Microsoft or within the companies that you collaborate with in training.

In our experience, I think we are much more leaning in the direction of using CareReds. So we have a lot of programs internally that are a mix, I think, in terms of tactics that’s important. Both kind of like, here’s a day -long training or a week -long training program, right? Which I think is really valuable. It gives you an opportunity to really dig in. But also really difficult. Difficult to find the time for. And so we actually have weekly tips. for how colleagues that are in similar roles are using Copilot, for example, internally to have more efficiency in their work. And I feel like that’s the kind of thing where, you know, is that skilling, is that training?

I don’t know, but it certainly is helpful because that’s the kind of thing that in my day -to -day job I can look to and integrate much more easily. And the other thing that we’ve started doing is hackathon -type exercises internally that are not just oriented towards engineering communities, but actually our corporate external legal affairs group, which is not just lawyers, but is a lot of lawyers, for example, having a hackathon that’s really meeting that community where we are and building a Copilot to serve our kind of day -to -day work. And so a lot of, like, different kind of carrot approaches is what we’re doing internally and where we see, I personally can say, like, I feel especially the latter two, it’s just hard to find, like, time to do a deep training program.

But if you integrate sort of into your day -to -day work, make it easy with these kind of carrots, you can really start seeing the impact, and that motivates you to use the technology more.

So, stick is out of the window, you can’t do that anymore, right? But we use carrots and budgets. Okay? When I say carrots, it’s basically appealing to the individual now and their glorification. So if it’s a patent, you’re filing it. The company doesn’t own it, you own it, right? If there’s a paper, you’re writing it. If you’re speaking at a symposium, you’re doing it, right? And that allows them to think. And then we’ve actually spent a lot of time through HR to try and explain that with the pace of change of technology, if you don’t upscale, you don’t change, you actually are facing extinction in about five, ten years’ time. Gone are the days where you can be there on the same technology for 30 years, will not work, right?

So we home in the message, provide that, and then provide the push. we glorify people that file patents, we glorify them within the company so that’s one. Second when I come to budgets, we actually leverage budgets with our segment heads. So they’re given budgets, they’re given training budgets, we also provide them headcount budgets and say can’t exceed. So we’ve been able to actually improve productivity with AI so we used to run on a utilization of productivity basis the metrics all service companies track at about 73 % five years ago. We’re already at 83 % and I think I can push this up another 2 % in terms of productivity levels in the company again leveraging AI and that’s the budget approach that we use but with the seniors.

So it’s a mix of both if I may to be able to manage this and motivate this. But it’s an ongoing exercise.

It’s fascinating maybe we’ll come back to it as we talk to a close. Let me shift gears a little bit and talk a bit more about security and trust and come to you Julian if I can. So I think we’ve recognized and we’ve heard it in different conversations this week that there’s a trust deficit around the use of AI, certainly in a public context. There is some fear, suspicion, and anxiety in a global context. I’m not talking just about India. YouGov carried out a survey in the U .S. last month, and in the context of fintech, they found that less than 20 % of Americans trust AI in financial services. And they’re also sort of struggling, I think, with some of the cybersecurity questions and issues, which you’re very well placed to address.

So if public trust in AI remains fragile and AI -specific cyber risks are growing, which they clearly are, what are the immediate steps that industry should prioritize to counter those threats? And… Things like prompt injection attacks. How can these solutions be scaled? Thank you. particularly for developing countries?

of seven. So other than the incentives that we’re giving you to learn these technologies, which of course is to the company’s benefit, it’s to your benefit because these skills that you’re learning and that you’re going to be using will translate to the next thing that you do, and it makes you that much better. If we do enough of that, not only are we helping the employees, but we’re helping the societies and the ecosystem that they live in, including in India. I wanted to add one additional area that we’re really focused on to address the kind of AI cyber threats, particularly relevant in India and other areas in the global south. I mentioned that one of the areas that we’re focused on is multilingual and multicultural AI capabilities, and one of the most important foundational reasons for doing that, of course, is that you have an AI that works well.

and in different languages and cultural context is reliable performs well. Another reason is also that AI that is not robust and it’s multilingual and multicultural capabilities does have additional security weaknesses. You mentioned prompt injection attacks and you know one way in which you can think about a prompt injection attack is basically if you have an AI system and you have the sort of safety system around that, someone who is misusing the technology can sort of try to break that safety system or get around it and one of the ways that attackers do that is by using languages that are not well supported in that model or system right so if a model or system is primarily prepared to perform well in high resource languages, but not in low resource languages.

Tamil, for example, or some other sort of language that is not really built in to how the model performs, if companies aren’t attuned to that, then an attacker could use that language and jailbreak the system, basically get around the safety system. And so it’s just another reason why it’s really important from our perspective, and we’re partnering with a lot of others in industry and government, so this comes back to a public -private partnership opportunity, to really work on multilingual and multicultural AI capabilities. One of the things that we announced this week is actually there’s a benchmark from an organization called ML Commons, which is a jailbreak benchmark. It’s actually measuring how robust systems are against that kind of prompt injection attack technique.

And we worked with a number of others to really build out the current version of that. which is really English -specific, to include multiple Indic languages and Asian languages in terms of its capability. It’s not going to solve the problem. It’s one step of what we see in the right direction. But I just want to draw that sort of really specific area of focus in India and other areas for thinking about the kind of AI and cyber threats.

That’s wonderful. Thank you.

Can I add a point?

Sure.

So this is about the rise in prominence of AI agents. And we have been constantly investing in self -defending systems, just like a human immune system. As agents grow and they can – the scale and speed at which they can attack infrastructure, the hospitals, the energy grids, we need agents on the other side. And this becomes AI versus AI story, where we are smartly inventing agents. And we believe, first time, with AI. We can reverse the defender’s dilemma. So the dilemma, many of you might already know, attackers have to find just one open wallet in this crowd, but defenders have to protect all the wallets all the time. And first time, AI will give us aggregate advantage to defenders because majority of defenders’ time, 80%, goes in drudgery and skunk work.

And AI can actually automate and uplift that work. So the entire stack of defenders can improve and uplift with AI. And we believe that we’ll be able to build a self -defending adaptive system which can protect us from various vulnerabilities.

Wonderful. Thank you. Well, we’re drawing towards the close of the session, and it’s been a very rich conversation. I just wanted to take a step back and ask you all, you’ve been – most of you have been here all week. And you’ve heard a whole host of different interventions and some very significant investments and initiatives. What are your conclusions? What’s changed? changed in your perspective when you look at AI for the future from your own vantage point? What’s this event given you a new perspective on or crystallized in your minds? Maybe, let me go back to Lee. Do you want to share your thoughts?

It’s reinforced for me, you know, something I’ve seen through a lot of my international work with OECD, with global partnership on AI, just the need for the global cooperation, and not just at the government level, but among all different types of stakeholders, you know, within academia, within industry, within civil society, and working together. And I think, you know, we can sort of pause at this moment and say, you know, if you look at the safety report, we’ve made a lot of progress over the last few years, but we need to continue to work together and not just focus on the harms and the risks that AI can have, but think about the benefits. You know, if we are able to leverage AI, we might be able to, you know, help achieve some of the UN Sustainable Development Goals.

I think one other thing I want to just kind of enter into the mix, you know, the customization of AI for different regions also depends upon data. And a lot of my work has focused on, you know, how do we create voluntary foundations so we can exchange data more easily? Like right now, we don’t have data standardization. So if I want to exchange my data with any of you, my data may be in a different format. As a former lawyer, a lot of my work is also focused on we don’t even have standard agreements. So if we want to exchange data, how can we easily transact and not have all that friction and transaction costs?

You know, we don’t have the Creative Commons licenses right now for data. And if we’re ever going to get to that localization and that ideal point where we’re customizing for different cultures, we’re going to have to have a lot of different tools. we’re going to have to figure out ways where we can voluntarily and responsibly share data. And this has been part of the discussion, but hearing the conversations over the past week kind of underscore the need to continue to advance that work while we work on some of the other topics that we’ve been discussing.

Great. Julian?

More than anything, what this week has taught me is I’m old and this industry is moving.

Okay, so stop saying you’re old. You don’t look old. You look great.

This industry is moving so quickly. Again, skills that are needed and considered to be important today will no longer be necessary in five years. And if the workforce and if the users of the technologies aren’t evolving with it, we all fall behind. So what is a great advantage and opportunity in using AI, the danger is it also cannot. obsolete at the same time. And we need to be very careful of that and how we use it and then how we help, hopefully, to promote this throughout the world in a way that makes it equitable for everyone.

Great. Thank you. Amit, Sachin, any reflections?

Yeah, I think one of my big takeaway from this week was some parts of the world are focused on AI as an influence. Some part of the world is focused on governance of the AI. I think India is focused on impact of AI at the grassroots level. Thinking about how AI will impact a farmer or a small school or an NGO or a small hospital has been the focus. And it resonates with me because mission of my team is to keep everyone safe at scale. And when I say everyone, it’s not just about Google or Alphabet or not just about our billions users. but the entire society, everyone at scale, and how to make sure we become the architect and not just the consumer of AI and make sure it reaches to the grassroots level is one area to think about.

I agree with that. So, of course, outside of the traffic bit, right? What you learn, if you ask me, in the whole week that I’ve seen is that if I, and I’ve been in this business for, I don’t want to date myself, so say a couple of decades and we leave it there. But people used to say India is a back office. That’s how it started in 90s. People said India is a back office. Y2K happened and they said the IT industry will be over, right? Because Y2K, that’s all there is. Today, the IT industry, engineering industry together is $600 billion. We move forward. People said, are you going to take data? And are you going to?

Is data going to get leaked? and then COVID came and India proved yet again there was not a single data leakage that happened from India Inc anywhere. There are some draconian rules. We don’t allow our employees to use USBs, blah, blah, blue, blue. Net result, zero data leakage, absolute privacy and the government comes down very heavily if they get something like this. So they’ve been able to create a safe environment. Move forward. People used to say is India a market? This last week and forget technology companies, if you just walk the floors, you see people like Schneider, you see people like Vertiv, you see others, they are developing products for India. In India, you’re developing products to the world from India and it’s no longer just a cost base.

So if I was to say there’s one thing that I’ve learned in the last week, it is that India is no longer the back office for AI. it is actually the front office for AI for the world and that’s the net summary that I would draw in the entire week that I’ve been here

Thank you, that’s very funny Bill

And I, you know, zooming out to the sort of highest level, one of the things that I really genuinely felt this week that has been very exciting to me is that there is a lot of energy around how to deploy this technology, how to have impact it’s been actually really fun to be in a lot of sessions with students and entrepreneurs that you can really feel the energy and I feel that it has the conversation around governance has come along and felt integrated in a really genuine way as well, if we look at the kind of summit series that kicked off a few years ago at Bletchley, I think it’s fair to say early on the emphasis of the conversation felt very safety and security heavy last year In France, there was a big pivot to trying to think about the opportunity.

And what I see in India this week is a genuine integration of those conversations and a deepening of those conversations. So really, what do we mean when we say impact? What really do we want to see in deploying this technology? And then sort of not taking for granted that, of course, governance actually has to come along with that. You have to really do the deep, hard work around things like multilingual AI. And there’s a real need for a partnership in moving those things forward. And there’s a real need to think about governance steps so that you can have trust in this technology. India actually just passing a law last week thinking about how to mark AI -generated content.

There’s a real sort of recognition that some of those steps are going to be important. And you don’t want to stop or have those steps sort of prevent deployment of the technology or realization of the benefits. But, like, you know, we have to do the deep work together to sort of move. Forward across a dime. A dime. and impact and governance together.

Thank you. Thanks, Amanda. We’ve got a few minutes. If anyone would like to chip in. Great. Hands are going up. The room’s filled, by the way, while we’ve been going along, and it’s been a great conversation. Let’s hand one or two mics out to colleagues around the room, if we can, to the lady here on the front.

Hello? Hello? Okay. Right. Thanks, and I appreciate the comments and the traffic. I think we’ve all got a traffic story. Now, I hear a lot of talk about upskilling, co -creation, which are all very important things. I agree. But what I’m also hearing a lot from, and I’m sure you all are too, is the issue of speed of this technology that could potentially outspace some of this real scenario. So my question to you is, you know, what do we – and this goes to anyone who might want to answer or has some real thoughts on it. What do you think might – be the gaps between that that we would need to address in a transition process between upscaling and real economic displacement.

Who can grab that? Yeah you put the mic Julian you’re gonna give it a go.

It’s a real problem right meaning technology is moving so quickly as I said years ago I would tell young people in technology learn to be the best programmer you can. Now with agentic AI especially with the usage of MCP where you can have multiple agents talking to each other sharing information it’s now learning to be the best user and prompter of the technology understanding the outcomes but there’s gonna be some displacement. It’s you know right now I would tell you AI especially in the security context I can probably eliminate 60 % of the things that humans have to look at today. but there’s still the 40 % where a human has to be involved to make a determination around risk to an entity, whether it’s a government, whether it’s defense, whether it’s a business.

And so it’s really helping them evolve to this next level of user, this next level of programmer, if you want to call it that. And there probably will be some displacement that we just can’t get around.

Gentleman in the front.

I actually have an extension of the same concern that the lady shared. The speed is one aspect, but also I think there’s a whole information arbitrage between the people who are creating and pioneering in the AI space versus the others to whom the information is reaching. And the impact of that on the power polarization and even the democracies. You know, that possibility I sense. And a lot of the conversation that I hear today is assuming that, you know, AI is moving linearly, but I see it moving exponentially. I agree. With a polarizing effect. Yes. Yes. Both. Both the polarizing effect and the effect, you know, like I think 40 % that Serge just spoke about. For me, that 40 % is not really 40%.

It’s just that we want to be very, very careful. But if we were to not care so much about how accurate and how much data standards we have. It could be 100%. You know, it’s very large. You know, I think the displacement can happen very fast. So I’m really concerned about how things are moving. I’m not sure if my concern is being shared by people in the panel.

Anyone want to respond?

I mean, I think we need to focus on AI literacy because, you know, again, the technology is moving so fast. How do we make sure people in their everyday lives? People in the workforce have access to education so they can continue to upskill. and I also think being in academia after having been in the private sector for we won’t go into how many decades but teaching students how to think I think a good student when you’re looking at your career trajectory it’s not just coming out of college with a set of skills but teaching them how to think, how to problem solve and I think it’s really the public -private partnerships that Amanda mentioned with academia is really important because a lot of times the tenured faculty, they don’t know how to teach that to students and bringing people in to tell them this is how you adapt, these are sort of what you’re going to expect in your career and I say this not only from the perspective of being in academia but having two children of my own in their 20s who are just starting their career and sort of expect the unexpected but learn how to be on your toes I think a lot of it is just having the good analytics skills, having good communication skills and if you have those core skills you’re going to be able to adapt and it will carry forward in the future.

Great. I think we’ve got time for one more question. Okay, gentle. Oh, the lady who, sorry, the lady who has the mic. She has the mic.

Thank you so much. My name is Rita Soni. I work with a company that’s operating in small -town India, delivering all these tech services that many of these companies are doing. And my question is actually for Amanda, because I think she was the only one who really brought up the digital divide that continues to exist, both in India and across the globe. I actually didn’t feel like I heard very much about how to actually bridge that. Yesterday I didn’t have one of those special passes to go to the events on the 19th, so instead I visited a local nonprofit called the Digital Empowerment Foundation, which has been around for more than 20 years, doing incredible work in rural India.

And they’re simply talking about last -mile Internet connectivity, let alone the enablement or ease. in the critical thinking that Lee just mentioned. So just a few more words on how it is that we can bridge this digital divide and make it more equitable, because I think the more folks are going to be excluded, the more different kinds of problems that we’re going to have.

Yeah, and I think you may have come in after we talked briefly about some of the work that we’re doing to address the digital divide. And for a lot of words, I would point you actually to we published a blog on Wednesday where we talked about investments in five areas that we’re thinking about to close the gaps that we see. And we actually point to the work that we’ve done using our own telemetry to sort of track these gaps and their trajectory and really lifted up our own concerns about the trajectory. And so among the areas of investment, infrastructure is really foundational. And we actually do talk in the blog about of course, infrastructure in terms of like AI.

compute capacity, but actually the fundamentals beyond, like, in terms of connectivity, energy access as really important as well. And then we talked about scaling multilingual and multicultural AI capabilities, really working with local communities on local use cases and the kind of deep work that we can do to sort of help bring the technology to people and see, like, even in agriculture, for example, we at Microsoft Research have done a lot of projects, like, in close collaboration with local communities and try to see, like, how could this serve you and then also learn from how the technology needs to evolve in order to do so better. And basically then also taking a step back and continuing to study diffusion so we understand, like, are our interventions working?

Are they not? If so, what can we learn and how can we improve how we’re intervening?

Okay, so time’s up, everyone. Thank you so much for your contributions and for joining us at different points during the conversation. Thanks to the panelists for a really rich and diverse conversation. It’s been a real pleasure to have you with us. And I think we end with a sense of optimism that no matter what the challenges of the digital divide and those other elements, there’s probably an AI solution to the AI challenges that we’re creating. Thanks. Thank you. Thank you.

Just a little background, why Ministry of Panchayati Raj exists at the centre, because rural local governance is a state subject. We are rather new in this business, we came into being in the year 2004. Our objective was, or the purpose why we exist, is how we can empower panchayats, how we can nudge states into having acts that really transform our people into self -governing, responsible local bodies and so on. So, as a part of our job, we also have oversight over how the ministry, how the panchayats spend their finance commission grants. Finance commission grants are devolution grants, they go directly to the people in their… bank accounts and then subsequently… all panchayats, all two and a half lakh of them they are present on eGram Swaraj right from planning to the payment stage, everything is done on a portal which is called eGram Swaraj this portal works in the English language so I will tell you in 2019 when we were starting something called the People’s Plan Campaign, I happened to attend a Gram Sabha in the state of Karnataka I was there for something like 45 minutes and I was felicitated and sat on stage and I didn’t understand a thing and then it struck me I had this thing that how do you expect these people really to relate to what is happening because it is public money Everybody in the panchayat needs to know what kind of plans are uploaded How many works got done that were asked for the plans How much did it cost them to do it And subsequently they can raise issues in the meetings pertaining to the works close to their residences And along came Bhashini I think we had in the year 2023 an event called Manthan Where we invited a lot of people from the industry to tell us how we could conduct our business better And so Bhashini was a revelation And imagine that a person from a panchayat is looking at the expenses page for his gram panchayat or her gram panchayat And then by the end of the month, he has to pay the expenses And by a click of a button, they are able to see it in their own language It was magic That was the starting point.

Yeah, and subsequently, of course, we went from there and we found out through a survey that what really hurts a panchayat secretary is not to be able to produce the minutes of meeting in time, which are very important, which are the only record of a panchayat’s proceedings. And then again, using Bhashani and another tool, we were able to create Sabha Sar, in which if you input the video slash audio recording of your meeting, you are able to get a minuted draft, which you can then edit and upload. So that was miracle number two. And briefly, if I could also address Swamitva, the scheme that you mentioned. The Swamitva is a scheme where. Drone surveys are carried out over all the village habitations, so there are these pictures.

that are subsequently converted to orthorectified images and they lead to property rights for the people living inside those villages. But the way the images have been captured, there is dense point cloud information, all of which was getting wasted. Why? Because we were confining our attention only to the orthorectified images. So we had the AI guys look at that and then they converted all those rooftops that they could see into the solarization potential. As a result of which now, out of the 3 .3 lakh gram panchayats where drone surveys have been carried out, in 2 .38 lakh gram panchayats, you can go to gram Manchitra, and you can zoom into your village and then you can click the icon corresponding to the solar ability potential and it will tell you roof -wise how many panels can you fit there.

We’ve gone further and we’ve integrated that with the PM Surigar Yojana portal. As a result of which, the Gram Panchayat can drive it like a campaign and lead to greater rewards for everybody all around.

Actually, it reaches the last mile citizen when you talk about those benefits. So India’s last mile operates in local languages and dialects, as you mentioned, solving that problem. So in your view, how critical is language AI in ensuring that digital governance platforms are inclusive and participatory and increases citizen trust and participation in Gram Sabhas?

Like I said, people are now able to follow what… what was something that was written in, they could still see it, of course. In the English language, then they’d have to go to the person who they knew to be very smart in the village and they’d have this person read it out to them. Now they can see it at their leisure. Not just people here, but people outside who are working in Mumbai can see what is happening in their panchayats close to Pune or something and immediately they can get active about it. And the militarization tool that I mentioned, that opens a whole new set of avenues now. You can have a record, then against that you can have action -taken reports and then you could have follow -up in the next meeting.

It makes it all amenable to a very systematic representation on portals. So that is what some of the states have already started doing. And it is truly remarkable that, anybody can go in there. And when I say anybody, I don’t mean just the panchayat secretaries. Anybody in a village can drill into their gram panchayat’s record and see that corresponding to the finance commission grants for any year, what was the plan against which how much has been executed, how many bills were prepared against each activity, and what is the status of the payment, whether it has been completed, where the asset exists, the geotags, and then you can zoom in and maybe see it on gram panchayat.

So there are great rewards for everybody all around. And we need to, of course, now intensify it through a capacity building training program. That is something we started doing from the previous year. But it has been an incredible journey. And it is being adopted all over.

So, Alokji, let’s talk about… Let’s talk a little bit about Sabha Saar Impact. let’s let our audience know about it and with its launch on 14th August 2025 MOPR introduced an AI enabled voice to text meeting summarization tool powered by Bhashini ASR services. So as of 4th February 2026 over 1 ,15 ,100 15 gram sabha meetings have been processed. So this is a good number I need a thank you for the round of applause. So what structural changes have you observed in the panchayat functioning after sabha sar?

Sabha sar was one thing that we carried out for the convenience of the panchayats and the panchayat secretaries as opposed to E. Gram Swaraj which was which was our selfish motive we wanted panchayats to plan there and show all their vouchers there so that we could tell that this is how the money has been spent but sabha sar actually came through and As a part of a survey that was carried out using RapidPro by UNICEF, we asked something like 8 ,000 panchayat secretaries all over the country that how do you spend your time? How much of it is spent in inspections and attending programs and meetings and making records? So one thing that came through was the conduct and recording of meetings, meetings was the, in 65 % of the respondents, that was the activity that was sitting, you know, very heavy on their entire time availability.

And so having realized this and having the help of Bhashani, we converted it into a tool. So in Bhashani, it’s very simple. There is no big. Standard operating procedure as it were. So if you’re standing having a meeting, there has to be a recording device. It could well be your mobile phone. And then through audio or video recording, you can just place it each time somebody speaks. And later on, you input this into the Sabasa tool. The Sabasa tool is not something that is a part of the device on which you carried out your recording. So the issue related to connectivity in villages is something that we’ve been able to sidestep. And once you do that, it gives you a draft minute of meeting.

So Bhashani turns it into English. And the English thing is monetized using the AI engine. Again, Bhashani gives it back to them in their own language. And that efficiency. Yes, it’s voila. The person can just make a few changes and upload it. And we have. We’ve had some heartfelt gratitude coming to us from villages. as a result of this.

Okay. So has the structured documentation improved transparency, participation tracking or monitoring of meeting frequency and agenda quality too?

Now that the minute is ready, if there are five items, ten items, so the states that have really gone ahead and adopted it, which is Odisha, which is Tamil Nadu, which is Tripura, all these people are into the second stages now where they are looking at the minutes of meeting and converting it into or refining it into tools that help them keep track of the activities after they have been created. We also realized through our meetings that why is the number just 1 ,15 ,000? So there are a whole lot of people whose languages do not exist on Bhashani. So from there, we ask those states to provide Bhashani with the necessary expertise so that they can train their bots.

And they’re already working on something like 11 more languages, which includes Assamese and Bodo and Maitali and Santal and whatnot. Yes. So those languages are also. So it’s been a very gratifying experience. And then the learning continues.

Yeah, it’s commendable that things have reached to that level. So over to you, Amitji, from an accountability lens, does structured documentation change behavior with the governance systems?

Thank you. So I think, you know, so if you have understood the enormity of the situation, right, what we are talking about, 200. 150 ,000 plus gram panchayats and different kind of languages. so just to circle back if you look at the frugality of the situation right so so so for example if you look at in india we generally people talk about either we live in a bullock cart stage right or we are aspiring for bullet train right so so the point is if ai has to tell us in terms of you know how we learn in the future how will we transform so we cannot i mean leave out 900 plus million people who are living in villages absolutely so the idea is not to make it very very urbanized you know very very kind of elitist idea that you know that ai is only for urban ai is only for industries ai is only for commercial sector so obviously this is a journey right so you have to start somewhere so for example i mean the frugality what i was talking about that we did not ask gram panchayat to invest anything right all they need to have is a lot of money and they have a mobile phone which any which way they have right and the idea is just to kind of record and upload obviously there will be some you know challenges and kind of resistance also in the beginning But, you know, once they get used to it, so for example, today we are asking them to kind of, you know, upload your recording, right?

The rest is done by system. And system also has a provision of, you know, human in the loop so that we can go and correct it. Now tomorrow we see the next step what we will be doing, what we can do perhaps, right? When the next meeting happens, we can also populate the agenda from last meeting, right? So what was discussed last time, what was committed, whether you are doing or not doing, right? And then everything goes to kind of public domain. So generally the people who live in city, they know that, you know, when there is a RWA meeting, nobody goes and attend, right? But they all, you know, wages, warfare in the WhatsApp group, right?

So same in the village also, it’s not easy to bring people, right? But once they start getting the hang of it, right? That okay, there is a meeting, I am getting the mom and it’s available in the public domain. We are using AI, AI is for good. AI can do it. AI can also be leveraged for rural sector, right? Why it has to be very, very elitist only for passport, say, wallet, say, right? Right. So so that’s just a beginning. It’s just a journey. Right. And also, if you see from an idea point of view, I mean, this is a phenomenal idea for Ministry of Panchayati Raj. Let me congratulate sir and the entire team to think of something like that.

Right. Because the AI is all about idea and use case. Right. If you have the right idea, you can do wonders. But you have to have idea and kind of, you know, muscles to execute it. So that way, I believe that in this whole documentation will do wonders for them. Graham Panchayat will also realize something which was missing in the most part of the word that, you know, the record keeping accountability, transparency, so on, so forth. Because generally these decisions were taken by some people only and executed by some. And the large population was largely kept out of it, knowingly or unknowingly. Right. So I think that’s what I said, that, you know, it will change the way they were.

it will change the way they think because this is only for a you know kind of we are starting only with a let’s say meetings but now they will start thinking and there will be demand from states and otherwise right what more can be done with AI so broader case would be achieved yeah Sabasa is an example like Praman we are doing we have launched this Pancham you know bought also for all elected and selected representatives so I think it’s a great you know kind of experience efficiency would obviously help them adopt I mean let me tell you in our own corporate meetings we are still some of us making notes right despite being on teams despite using co -pilots despite having all tools at our disposal but we are still using it right we expect a junior guy to take notes and circle back so that’s a cultural change which you have to also see and these changes and these changes couldn’t have been possible if we wouldn’t have the infrastructure like Bhashni right because ministry on its own how ministry got benefited, we have infrastructure like Basni, right?

We have the you know, GPUs got available to us through the NDIA mission, right? Otherwise, you know, procurement itself could have been a big challenge, right? So we have a team to kind of build applications. So I think you know, it takes a village to move something, right? So that’s what has happened here.

Thank you for sharing your thoughts. In fact, just continuing with that, the Department of Drinking Water and Sanitation has actually approached us that the meetings of their village VWCs village water committees they want to use Bhashini for that and there has been some initial interaction between the two.

That’s commendable, I would say. That’s awesome. So Alokji, let’s talk of some implementation challenges in rural India with AI. AI in rural governance is transformating but complex. So what are the biggest operational challenges, infrastructure, though a bit, I think Amit you were about to share that, but then infrastructure, training, dialect diversivity and connectivity. So what challenges are you facing? How receptive are panchayat functionalities and rural citizens to AI -enabled systems?

Challenges, of course, there are many and you would have anybody tell you. What we have found out, the adoption of eGram Swaraj by our villages is gram panchayats. A case in point, Uttar Pradesh has got something like 59 ,000 gram panchayats. And for Uttar Pradesh to onboard eGram Swaraj seemed like an impossible task because it involved registering your digital signing certificates and then everybody agreeing to completely dispense with checkbooks. All your payments were then going to be… Can you imagine Uttar Pradesh did it in 40 days flat, all 59 ,000 gram panchayats. So my point was that if you are ready with a product that addresses their needs and it is friendly and it meets, of course, my need was that I needed the money well accounted for and their need was a system that could make it very easy for them to do it.

So we met halfway and if UP can do it in 59 ,000, I am not prepared to hear an excuse from any other state in the country. It’s a trial by fire. Likewise for Sabasar. Sabasar, again, I said initially that there was a demand that was indicated from the state. So when we set out to meet that, we were clear what is it that we are looking for and people were so forthcoming. In fact, Bhashini also enabled me. to write letters to the states in their languages and people were gushing with affection and what not. I got a letter in Telugu for the first time and all that. So there are challenges but then the Ram Panchayats are predisposed to meet you halfway.

So you need to begin that journey and we have seen that with regard to a number of things. There have been campaigns every year they carry out a campaign from 2nd October to the 31st of December which extends to January typically where all two and a half lakh Ram Panchayats prepare their Ram Panchayat development plans and upload it on the portal. So 2 .5, 250 ,000 Ram Panchayats all of them planning for the next year and so before you enter the next financial year their plans are ready. I mean we don’t… We don’t do it in the departments, in the ministries. And all these Ram Panchayats have… not done it once, twice. They started in 2018. They’ve continued to do it ever since.

In the COVID year, there was a request that campaign. So there was a massive pushback from the states that no, we want to do it. The inertia was so great that they still did it. So there are challenges but if we make an application like you were saying that this is a simple recording device, this is a mobile phone, there aren’t things that you need to procure to set it up. So if you make a simple tool, people would grab it with both hands. So I think that is the embracing of challenges rather with the response we are getting with Bhashani.

So for ministries delivering last mile services such as Ministry of Rural Development and the Ministry of Agriculture and Farmers Welfare, what lessons from MOPR’s AI journey would you share? How important is open architecture and in your sense?

That is dangerous territory. I am not in a position where I could start advising anybody because they’ve got pretty robust systems of their own. If you look at Manrega Soft and the PM Avas Yojana, because they are running schemes which are very pointed. Avas Yojana is just about houses. Manrega is a scheme where there is, of course, it’s as large as the things that you do in the Finance Commission grants, but it is fairly well organized. And in all of these, typically, the beneficiary is the individual. In Panchayati Raj, there are individuals at the end of it, but our emphasis is on the institution, the panchayat, and not just E. Gram Swaraj and the things that we do for their accounting and planning.

We also hooked up with the… Meteorological Department… and there are daily forecasts being generated for every gram panchayat. This people are able to see on their phones and all with the similar ability as they are able to see everything using Bhashini. So it’s a great enablement all around and it can only get better.

Absolutely. So, Amitji, over to you. How critical is open architecture ensuring long -term sustainability and avoiding vendor lock -in?

If I can take a minute and talk about the previous question also.

Sure, please go ahead.

Sir rightly mentioned that different ministries have got a different mandate. It’s not an apple to apple comparison. But see, you also have to see the panchayati raj, the main role of panchayati raj, what I understand is a mobilization. because they are not running major schemes on their own compared to others. And generally the best practices doesn’t have to be in form of technology or architecture only. The idea is that if you go down from top, there are two different ministries and if you go to the village, you will see the same infrastructure, same set of people are only working from both departments. So the idea is if one can do, others can also do. So there is a lot of learning in terms of method that how we could overcome, how could we mobilize, how we could implement some of these solutions.

And I am sure we know that RD and agriculture are also doing a lot of things, but their mandate is much bigger. But they can also take a lot of pride or learning from the success which we have. What was the second question? The second one was that how? Critical is open architecture in ensuring long -term sustainability and avoiding the window. So you must be hearing this word called sovereignty quite a lot, right, nowadays. So the whole idea of, you know, being sovereign in any part of the, you know, technology, be it defense, be it IT, be it any way, is the survivability, right? So the idea is despite, in spite any kind of, you know, geopolitical risk, we should survive.

Our system should run, right? So for that, generally people confuse sovereignty with also making India local, et cetera. So that’s not the case, right? We will always have some technology from outside. But we have to design in a way that it is kind of ready to shift, right? So either from a technology point of view, we have the interoperability, the standards which we have chosen, the models which we have chosen, the infrastructure which we can move around, and the teams which we can control, right? Right. So the data residency has to be within India and data is with us. So obviously if we have trained on one, we can train on another, something else also.

So the idea is also to look a little bit long term. See, what has happened that when we started, obviously, there were a lot of POCs. Nobody knew, right, how AI will behave. Still, we don’t know. Still, we don’t know, right? I mean, so obviously, that you have to start somewhere, right? And then you have to also ensure that in future, when we start with one use case, it becomes easy, right? When the department itself becomes fully AI enabled and we have 10 AI use cases running, then it becomes a problem, right? Problem of management. So that’s where I think we need to plan better for future so that, you know, we plan. I mean, it’s not that a use case is defined, then we found an easy method of procurement of infra or the model which I knew.

So going forward, I think there will be a platform approach, right? So where we have to think for future also that, okay, these AI cases are likely to come in future as well. Different kind of AI, right? DL. DL. DL. DL. DL. DL. DL. DL. DL. DL. DL. DL. DL. and accordingly we have to have open architecture like the way we did in a normal digital transformation. Even digital transformation, there used to be time where we created our own independent monolith applications. But now we are creating applications, you know, which are more API -based, can integrate with anybody, right? And futuristic, can scale our modular. So same concepts have to be used for AI

Well said. So I think adoption comes with responsibility and that’s what you are scaling at, looking at. Swalokhji, Sabha sir demonstrates how language AI can power grassroots governance. After Sabha sir’s success, what deeper integrations do you envision with Bhashini and what does the next phase of collaboration looks like? Let’s talk about that.

About 16 have already started providing all those common minimum services. So minimum se nahi chalega. We wanted more. So now we had like a model list, union set, if you will, of all the desirable services that were being delivered. And the ministry carried out an exercise through an expert committee. And we have a much bigger list now. So we are not satisfied with the minimum. Now we are working towards that. But I think that AI has great potential in helping us. Thank you. So service delivery is something people don’t know to expect. and we would like through and people are going to be speaking in any number of languages. I think the next step, my government is something that has always been very invested in providing services to making ease of living easier as it were and providing all manner of things.

Everything is finally a service. You need to look at a doctor. You need your road fixed. You need a street light to be working. You want the log water to be drained or something. She needs more attention than us. Okay, over to you. So people should come to expect. they should demand these services from their Gram Panchayats. There are mechanisms of doing that because Gram Panchayats don’t have a lot of resources in terms of manpower, in terms of people who are at their beck and call to carry out the activities that are flowing from the Charter. So there are systems in a lot of these villages. You have common service centres in some states. They have their own system of common service centres like UP, like Bapuji Seva Kendra in Karnataka, like Me Seva.

So we need to take that further and we need people to be able to talk and find out if a certain service that is available to them, can they avail it in their village? If they are to do that, what is the mechanism? And if they’ve already made an application, that what should be able to tell them that where that thing currently stands? so that is a very wide area like I said that there are a number of services we also learnt of a pilot that was carried out in Guwahati where the bus used to have a camera it used to drive through, capture all number of images and basis that it would assign issue labels to them as it were if there is a drain overflowing so it takes note of that if there is a pothole then it takes note of that and then it assigns it to all these agencies whose job it becomes now to fix that so not that but maybe we have a mobile interface called Meri Panchayat which ports a lot of information from E Gram Swaraj Meri Panchayat also has the capability of capturing images of the issue that is being reported I think the next step is that image it makes sense of the image and it assigns it to the necessary department.

There are people who are mapped whose job it is to carry it out and within a certain amount of time it doesn’t happen, then there is escalation. We need to go deeper into that system. That, I think, is the next frontier. And, of course, because it involves vocalization of your demands, so bhajani is absolutely critical in this. So when we say there is a long way to go, I think that phrase is no more relevant. It’s a short way, but not even a big journey, an intelligent journey to move

So India is building public digital infrastructure for AI at scale. So how do we balance scale with accountability and public trust? We have talked much about how we are building things. But let’s talk about the other side. And can India lead the world in population scale?

Of course it can. I am sure about that. But then multilingual AI for governance, when it comes, you would like to have a short -ended first. So one thing you all have to realize that whatever we do is a population scale and unparallel because of our size. So even our POCs exceed the kind of performance of European countries Our UP sir talked about UP 60 ,000 panchayats If you look at UP maybe it will be in top 10 country in terms of population and size. I think the world is vouching for us when it comes to the use cases So see if you look at that we have got that scale now. We have the experience behind us We did Aadhaar, we did UPI, we did Fastag, we did GST and we did Income Tax.

So now we have that confidence behind us that we can do anything of scale and with the same Prugal approach we will do 10 times cheaper than Western world and certainly not worse, better only. and also from last decade we have evolved right so for example the concept of privacy like dpdp act consent based usage like you know adhar brought so a lot of things have improved from a policy side of it now now once you have policies in place systems are easy because system themselves act as a rule you know once you have policies in place then you don’t need so much of human intervention or discretion so since we have done it since we have kind of you know done so much so now if you look at the very simple case bhashni i remember four five years back and i and amitabh used to i mean kind of debate also whether we need a bhashni okay right because we we had some of the google translate services so on for forth right but the idea is that i mean in the hindsight that was the right call right in future we have to have something called sovereignty word right we have we don’t have to depend on I mean we need to be frugal and we don’t want to use you know the applications which are very expensive from a taxpayer money point of view so similar things we have done a lot right so I think the next step for example if you look at roam around in AI summit you will see how many LLMs and SLMs we are building on our own right honorable ministers talked about five layers application I think we have ample talent to build applications LLMs we use open LLMs but we are developing our own and Basni also like one of the you know common infrastructure energy will take care right infra and chips anyway will have dependency but that’s the rest of the world also has a dependency right not that everybody has a rare earth and everybody is building chips so that way I believe that you know that and because we have that technical know how also I mean that’s our kind of bread and bread and butter nowadays right so we’ll be able to take the learnings from all these systems and we’ll move forward as of now we were a bit slow in last year or two because AI itself was new for everyone so we took some time but now I think from this year onwards we’ll really scale it up because we have tested the blood, we have seen the success and we will

sure, thank you for sharing that so as we come towards the closure of this conversation I would like to leave with the one final thought which is like if Panchayati Raj institutions are the foundation of democracy can AI when built on a public stack and powered by language inclusion become the strongest enabler of participatory governance in 21st century just closing thoughts from you both Alokji, would you

absolutely he was just telling you that that we’ve been able to do things at scale this thing about UP that I told you I wear it like a badge that to have done it in some place so and it’s not an easy ask because there are so many stakeholders they’ve got various kinds of issues of their own you’ve got to engage with them address those things and if my problem is well defined and if I know what kind of a thing is going to help me redress that like Bhashini did for us I think that what you said is going to come true because that is so being able to understand my problem and knowing what parts of the problem can be fixed in what manner using the various tools that are available that is the key and I it’s not an over simplification but good servant bad master so that is something that stays and it is not it’s not going to land you in the right places if you just let it go around like an animal.

But then if you know where to put it, what modules to be inserted, where, what has been used in the background. And so that would make you more confident. I’m not really an AI person, so I’m just speaking on the strength of what I have learned. And the experience thus far has been outstanding, partly because we’ve had a very good partner. But other than that, I am not, you know, I’m not throwing it all open out to AI. I don’t wear T -shirts saying I love AI or something. But I have a problem, and it needs fixing. And I need to be able to know what aspects of AI can help me fix that in the best possible manner.

And that’s my thing on

So like Sir said, you know, Sir is not an AI person, neither am I. So if you look at, you know, that… But he was transparent enough to share that. No, no. So look at that way that none of us were, right? Yes, exactly. Because if you’re talking about AI, I have been doing this, you know, digital transformation for public sector for over 20 years. Yeah. Obviously, there was no AI, even there was no DPI, DPG also, you know, what we kind of retrofitted with the names, right? Right. So if you look at the idea of Panchayati Raj itself is a participative governance, right? That people have to assemble in the Gram Sabha and decide on the money which they’re getting, how to spend and prioritize.

Absolutely. And if AI tools like Praman and Sabha Sar and, you know, Pancham can help that strengthen, what best, you know, you can expect from a participative government, from a democratization point of view. Yes. So I think this sometimes, you know, that technology becomes secondary. Yes. In my view, most of the time, right? The ideas have to be clear in terms of what you want to achieve. and what problem you want to solve, what scale you want to solve, what are the guardrails you have to kind of, you know, also put in place. So, for example, when we do AI, that it cannot be 100 % autonomous, right? Of course. And it cannot be 100 % human in the loop also.

Because if we have each and every transaction being, you know, approved by human in the loop, then it defeats the purpose of AI. And there is no AI, right? Then we are still living in the rule -based algorithms. Algorithms. So the idea of, you know, that AI will be that we also train, monitor, have the mechanism to take complaints, have the mechanism to perfectly, you know, kind of train it better so that we improve our accuracy. So that is how AI journey. So AI journey is slightly different from the previous digital transformation journey, which were more like a transactional systems, right? So that way, I think, if you look at currently also Sabasar, I think whatever I am hearing from people market teams also, So it is giving great accuracy, right, in terms of translation and summarization.

And I’m sure whatever there are little bit areas to improve, it will improve on its own. So we cannot stop it, right? So once we have boarded a flight, then we can only get down at where we have to, right? So I think future is bright. And also from a MOPR experience point of view, it will also, I’m sure, energize and motivate a lot many others. I can say with my experience that if MOPR can do in rural, we can use AI tools. There is no stopping for us as a nation. This is truly an achievement when it comes to MOPR with the government.

So you want to say anything regarding this, Alokji?

I thought of another application that works. That is something we’ve been working on, which was spatial development plans. Okay. we again engaged with a lot of panchayats that were close to the highways okay so typically if a panchayat is on a national highway close to a big city and have a population of 10 000 plus then you were eligible to participate in this program okay so there were 34 gp that we involved and we got the planning and architecture colleges to prepare spatial plans for them spatial plan would be futuristic it would zone and it would you know assign it would look into the future and see how this place was going to grow it would devise road networks or something and tell people what they would become over a period of time we had a conference with the with gram panchayats around bhopal bill and the people were so annoyed We don’t need a spatial plan.

Over a period of time, of course, we told them what it was going to be, but we had this epiphany that people need to be able to see what this spatial plan will help them become. And then we went into the next national conference. We had for each of these 34 spatial development plans a visualization. And we showed people that if you want to become this, you have to do this. And then there was greater enthusiasm. So the people on whom this plan is, who are going to be subjected to this plan, if I could use those words. So these people, if they’re not on board, there is no way you can carry it out. And that, I think, is wide open.

And we’ve had after that. But the entire state of Andhra Pradesh has gone ahead and said that all their planning is going to be spatial plans. So that is something that is amenable to AI tools. And a final thing that I remembered that lots of times we need to convey through audio video messages. He mentioned Pancham. So Pancham is a WhatsApp -based chatbot platform which allows us to have two -way conversation with all the sarpanchas and panchayat secretaries in the country. So all these people. And so if there is messaging that needs to be conveyed, if there are videos that need to be quickly created using AI tools, that is something that would be hugely effective in getting the message across in the quickest possible way.

Thank you.

Thank you so much for such endeavor. insights on the Grampanchayath and how things are working behind. Actually, I’m sure the audience was truly, they were unknown about what’s happening around and this conversation has given a new tangent to how we look at the rural development. Thank you so much Shri Alok and thank you so much Shri Amin for sharing these thoughts on Grampanchayath development. Thank you so much for this fireside chat. Thank you. I would like to call Ms. Deepika. to please felicitate Mr. Alok.

models or talent, how we can ensure that it works in a federated manner. I think I’ll just, I was discussing and maybe I’ll just focus on one piece, which is on AI -ready data, if I can focus on that and leave it for the esteemed panelists on the large number of issues. Some of the elements that we are focusing on include, one is on how to make it more discoverable. That would be a very basic point to ensure that it’s discoverable. Second is how to ensure that the data sets are trustworthy, and that would be the second element. The third would be on the interoperability, and the fourth would be on the usability across systems.

So on discoverability. On discoverability, the metadata of that structure is extremely important, so that would help to, that’s a first. element of having a metadata structure which is understandable and well defined and can be used across the second on the trustworthy part would be the quality assessment so we’ve developed as kind of a quality assessment framework which focuses on the quality of the data so that to ensure credibility on the data interoperability a lock would depend on whether data can talk to each other what is the unique identifiers that we have which will ensure that the different data sets whether are they talking about the same thing or different we are able to identify that and the fourth would be on the usability across systems would be based on the standards and classifications that we have whether it’s a common definitions and common standards so that two sets of data don’t refer to the same thing and I suppose this really forms the bedrock of making a data AI ready and that That’s something that we’re working with ministries and governments and state governments across the country.

And given the importance of data sets in the AI infrastructure, it has an important part to play. The other aspect on data is also on its dissemination and access, on how we are able to ensure that data sets in themselves have value beyond AI and what kind of dissemination and mechanisms can be there which will make it usable for people to leverage them for business while preserving the privacy aspects of individual data. One other thing, since we are talking about and the panel will be having discussions on AI infrastructure, I just wanted to focus on one thing that I think discussed, over the past couple of days. has also come up that the existing models seem to be extremely data infrastructure, infrastructure heavy, whether compute infrastructure, data infrastructure.

And every time a new query is put out to the model, is it necessary for the billions of bytes to be again run through again and the gigawatts of power that we need? And are there alternative mechanisms available? And I just want to highlight yesterday one comment which stays with me, is what Vishal Sikka had made, that when we talk in terms of AI infrastructure, we talk in terms of gigawatts of power. Compared to that, a human being requires 2 ,000 calories, which is only 100 watts. So are we missing something out there in the infrastructure? And perhaps a greater focus on the models going forward is there. So I’ll stop here. Thank you for inviting me. Thank you.

Thank you so much, Secretary. And I’m now going to join the fireside chat here. The discussion that we have planned will cover various different aspects of digital infrastructure. So when you hear digital infrastructure, you might be first thinking of the data centers and the compute. But we actually want to have a conversation that also encompasses the solution side, the skill side, so that we really look at the whole spectrum of infrastructure, even standards. So these three S were introduced yesterday by ITU’s head, the three S of solutions, standards, and skills. Kind of a nice way to open up to the panel. We have different perspectives here today. And we’re going to try to stick to time.

But let me introduce you to this panel by asking the first question. And I would request that each of the panelists then quickly states their name and their institution to shorten the time. What we would like to hear from each of you is that from your vantage point, what do you see as the most critical gap, the most exciting opportunity for the global south in generating positive impact through AI? So we’ve asked each of them to think about a concern or opportunity and then also to maybe link it to strategy or vision. So maybe I’ll first go to the lady on my right from the WTO. May I request you for your perspective on the big challenge or opportunity?

Thank you so much to the Asian Development Bank for the invitation and the organization to this interesting conversation. My name is Johanna Hill. And I… I am with the World Trade Organization. And let me start out with the opportunity side of the equation. We really are seeing that AI and trade, when they work together, can offer important opportunities for developing countries and low -income economies. Our projections at the Secretariat have led us to believe that by the year 2040, trade could grow by almost 40%. So that would be the 40 by 40 effect. But then here come the caveats, right? For that to happen, for those opportunities to really be realized, one element that is really important is the digital infrastructure, the skills that you mentioned, and policy readiness.

You know, we’ve heard throughout this conference and before the important opportunities and applications in different sectors, in agriculture, health care, new services being developed as we speak, new services and goods that are becoming more AI -related, more tradable, and we are also seeing that that can have important opportunities. for the smaller firms in developing economies and in the big economies also. We did a survey with the ICC that we published last year on the opportunities for businesses and for small and medium enterprises. And of those respondents, many of them were saying that they’re already using AI, of course, from bigger companies, more developed economies. But even the smaller firms are also seeing opportunities in areas like market intelligence.

So we do see that it can be a game changer.

Fantastic. And one of the things that I’ve been hearing a lot at this summit is that specifically the SMEs, the technology has moved so fast that there’s a huge adoption gap and understanding of how they can actually integrate the AI into their business models, into their little shop that takes a picture of a product and uploads it quickly. AI can be super helpful in this but hasn’t yet reached that audience. Maybe I’ll turn it over to you. Maybe I’ll turn to the other side and request our colleague from Uzbekistan to share his perspective.

Thanks for the question. Thanks for having me here. Let me talk about the gaps which exist in our country. I think the first one is unequal access to compute capacities and I think advanced AI digital skills. So in that sense, these foundations play a crucial role because if you don’t bridge those gaps, many countries, nations will be just the consumers of AI rather than creators of AI value. So in that sense, Uzbekistan is advancing strategic ideas. First one is developing human skills. So in that sense, Uzbekistan is advancing strategic ideas. First one is developing human skills. So in that sense, Uzbekistan is advancing strategic ideas. So in that sense, Uzbekistan is advancing strategic ideas. So in that sense, Uzbekistan is advancing strategic ideas.

So in that sense, Uzbekistan is advancing strategic ideas. So in that sense, Uzbekistan is advancing strategic ideas. So in that sense, Uzbekistan is advancing strategic ideas. So in that sense, Uzbekistan is advancing strategic ideas. So in that sense, Uzbekistan is advancing strategic ideas. So in that sense, Uzbekistan is advancing strategic ideas. So in that sense, Uzbekistan is advancing strategic ideas. So in that sense, Uzbekistan is advancing strategic ideas. So in that sense, Uzbekistan is advancing strategic ideas. So in that sense, Uzbekistan is advancing strategic ideas. So in that sense, Uzbekistan is advancing strategic ideas. all stratus of our nation, starting from students, professionals, and public servants. So we are not concentrating on the tech sector, but also we try to cover all the spheres of our nation.

And secondly, we are developing our infrastructure. For that reason, our government is allocating around 200 million USD. So to create our own government data center with supercomputers, GPUs, acquiring from NVIDIA. And also we are working with DataVault, a Saudi Arabian company, to create an energy -efficient data center, which is based on renewable energy. It is a very big project. It’s around 5 billion USD. And the data center will be… put into operation within two, three years. hopefully and also we are trying to develop our government strategy we adopted a strategy 2030 last year and this by by the year to 2030 we’re trying to get there early reach the export of AI related products by five 1 .5 billion USD so

fantastic so either by coincidence or planning you touched on the 3s the solutions the skills and the standards the policies fantastic thank you so a very comprehensive view with multi -pronged strategy that you didn’t introduce yourself so I just say you with the Ministry of Digital Technology and an institution quite focus the center of the development of AI and the digital economy Fantastic. Okay, let me turn back to this side. So from Indonesia, we have someone who’s actually in this skills domain. Would you like to share with us what you, from your vantage point, perceive as the key opportunity or challenge?

All right, thank you. Hi, everyone. I am Professor Hamam Riza. I am the co -chair of the National AI Roadmap Indonesia 2030. And also I am the president of the Collaborative Research and Industrial Innovation in Artificial Intelligence, the organization that was founded in 2020 when we launched our first national AI strategy towards Indonesia 2045. That is the vision. And I think AI will take us there, really. So from my vantage point, I think… We are… we have no we need to move beyond numbers even though we understood that AI economy will create millions of jobs and also potential economy of up to 1 trillion so from my vantage point from the Indonesian perspective global south they are basically triple deficit in terms of what we are going to it is the most challenging one the first one is that certainly about the data and infrastructures, the compute infrastructures we are still lacking the connectivity the networks but as I have marked down here in order for us to smooth up all the AI use cases for public services for health services, for agriculture, and many other things, you need to basically solve this triple deficit.

And that is also regarding how you need to develop the AI talents. There is a significant lack and scarcity of high -quality localized data centers tailored to Indonesia, as well as shortage of AI skilled talents that limits the capacity of long -term innovations capabilities. So our government is addressing these gaps through the national roadmap that I co -chaired. And our primary concern is how the digital divide and how the AI divide, which is the digital divide. Which is created by this. generatively I didn’t think I towards many of the public sectors in Indonesia in an in general in the global self that we can tackle so that while there is nine to two percent of our skill knowledge worker but we are still using you know a very basic AI tools and needs to be aware of all the and all the risks you know applied to the output of this AI tools so those those things are that I think will be my point of view towards closing the gaps for the global south and especially for Indonesia

thank you so much and it’s of course one of the most populous countries in Southeast Asia It’s a very young workforce. 270 million. Yes, and startup buzz in Indonesia is also palpable, so lots of potential in Indonesia. Last but not least, I want to go to our Asian Development Bank Country Director for India. Mio, can I request you to share your views?

Thank you. From ADB’s perspective, of course, foundation is important. We need to have a power supply, stable power supply, and the devices that people have access to, and reliable broadband, even in our office. So that’s a foundation. But we are in India. Do we expect India to put so much money on foundation to have a ground -level impact? But India has a scale. So what we need to focus is, as others already said, is a service. So we work on agriculture sector, water supply, and even irrigation sector where AI is widely applied. Because of the scale of the people that we have in the global south, while we work on the foundational infrastructure, at the same time we really have to work on how AI can be applied at the service level.

And this is where ADB would like to support. Thanks.

Thank you, Mio. So we have, as you can see, different perspectives at the same space of how do we get at grappling with this massive development opportunity that AI represents. For this first round of questions after the opening, let’s go into the foundations a little bit. I want to go to Uzbekistan again. And you already mentioned you have… ambitions on infrastructure, policy and skills. Now, how do you actually balance this in terms of priority setting? Can you go for all of them all at once? How do you finance it? Does this keep you up at night, how you balance these three different strategic objectives?

It’s a tough question because we are a developing nation and money is always a scarce element for us. So anyway, our government is trying to allocate enough resources so we can cover all the aspects of AI development to create the AI ecosystem. First one, as I already mentioned, it’s a strategy. 2030, which sets our priorities, which is human -centered AI. And secondly, with the government trying to allocate enough resources overall now the government announced about 300 million USD for development of AI and the money goes to first of all implementing projects in the government sector in the social sphere healthcare, education, transportation cyber security and etc and also government is trying to provide necessary infrastructure building data centers acquiring GPUs and also we are now creating a data lake which will be collecting the data of the government sector so SMEs, startups and other who wants the data they can use those data for free or for some money usually free and anyway we’re trying to work with other countries as I already said that we are we have a good project five million AI leaders so United Arab Emirates they helping us to helped us to build this program and it was launched now over 1 million people already registered and go to training certifications there so also we are trying to attract foreign foreign investments and now government announced very good in tax incentives and other incentives to for example if you are want to invest in a you know in Uzbekistan and try want to build a data center which costs over 100 million USD you will get very cheap and take the intensive and customs exams and etc so going trying to balance with cautiously but still providing necessary conditions for the to build a ecosystem

very impressive and since I had the opportunity to chat with somebody else from Uzbekistan this week I also know that in your KPI as a public servant AI roll it out has entered that KPI space so that’s always going to make a difference let’s go back to Indonesia now I’m gonna you know of course skills is your comfort zone but can I ask you about the infrastructure side I know that the hyperscalers the big cloud companies international companies have invested significantly in Indonesia now how do you see that now moving into the AI age and is that a big step forward for you? is there a lot of activity on additional infrastructure build out what do you see happening in that

yes so I see these questions and I’m really eager to answer this because suddenly our infrastructure is undergoing a transformation really to meet the demands of the AI demand and certainly with the ability of many of these new infrastructures coming out of the government and also from the business I think benchmarking with many other countries including you know in the regional ASEAN take for example the presence of the global hyperscaler in the country have established actually multiple cloud regions in Indonesia. But certainly this needs to be amplified because as you know 10 years in many other technologies, one year in AI, right? That’s what they are saying. So how do you can fulfill this demand of AI compute massive data for training because you need to build up our own for example large language model that can align with our cultures.

So those hyperscalers needs to move beyond just being a single a host for this you know many of the AI models from outside of the country right so and the infrastructure readiness is also being federated by our chief toward the sovereign AI we are now preparing the presidential regulation actually to push forward the innovations the investment and we need to collaborate with many of the hyperscalers and we are ensuring that the physical infrastructures like the GPU data center and localized edge computing yeah is going to be present in the country. And one thing that the Vice Minister of Communication and Digital Affairs mentioned to me yesterday that we are struggling building up the ecosystem. That means there will be special economic zone for these hyperscalers and new data centers being brought forward in order to align and be part of our national AI roadmap, AI journey in Indonesia.

And we are going to prepare ourselves in this AI transformation so that our data… digital consumer… is going to be part of our transformation. The technology is accessible for all. Even, you know, what we are right now, you know, participating in the India AI Summit says about democratizing AI for all. So I think that is a very significant theme that is also part of our national AI roadmap. Thank you.

Fascinating. And, again, I think you as a large economy, you have that opportunity similar to how India is also portraying it this week of really wanting to develop your own, you know, language models and really playing in that league. However, there are many countries. also countries we work with who don’t have that kind of scale and who need to look at it quite differently. So the different nuanced strategy that you mentioned of investing into the big AI, the small AI, the edge AI, all these different pieces, very interesting. With this dynamic, can I turn to WTO? How do you see trade competitiveness evolve? That’s really your space where you are at. What are those interesting approaches that are emerging which could help support maybe the cross -border collaboration while you also, of course, respect data sovereignty?

Countries will need to collaborate, right? There’s not enough money to go around for everybody to play in that top league. So trade competitiveness, what do you see there?

So I was talking about the opportunities of trade growing by the use of AI. Okay. And if you think about… That growth comes from the lowering of trade costs. It comes from powering AI -enabled goods and services crossing borders. And it’s… Also, new products and services that are going to be invented are being invented by AI. And when you talk to business, when we asked through the survey, some of the constraints that they are having and doubts in the use of AI have to do with competing regulations and having a high cost in trying to comply. And fragmentation is actually an area of data, for example, that can become a problem. And so we developed and published last year in the World Trade Report what the Secretariat calls the AI Trade Policy Openness Index to help regions measure how they’re doing in that space.

And in there, you can see, for example, that some of the lower income economies can seem quite open in that space. But it might be because of the lack of regulation. And when you talk about AI, I think what a lot of countries and customers are saying is that AI is not a good thing. What customers are looking for is, you know, it’s AI. that is responsible, you know, trading AI with trust. So just not having regulation can also be a disadvantage to your competitiveness. So starting to look at those things that way. And then in the part of the solution side, definitely the regional approaches are important, those collaborations, and sharing infrastructure, for example.

When you don’t have those economies of scale, those huge investments come in your way. And then not every single company or every single country is looking to be on the edge of things necessarily, but we do want to adopt AI to boost our economy and our competitiveness.

Well, thank you so much. And I don’t know whether people heard about this new initiative that the working group on the democratization of AI, of compute, has come up with. ADB is actually supporting that. Really, this is… This has not yet evolved, right? This collaboration on the infrastructure. How do you share that properly across borders? It’s still new territory and very interesting to see. Can I turn to my colleague, Mio, and request her to talk a little bit about the engagement that we’ve had with member countries. What does demand to ADB actually look like in this space?

try to invest in the township planning and the implementation. Also, we can have a water supply road project that can be connected to the industrial parks so that private sector can invest in the digital -related facilities. So mobilization of private capital is one. And the second is it’s an application across sectors. We just don’t look at the single sector project. As I said, we can work on road and water at the same time. And while we work on the Agri -AI project, we work with the building capacity of that institution as well so that they can handle the AI. And the third is the knowledge. As I said, we support quite a bit of this master planning or the strategy development at the municipality or the state or even at the regional level.

We see India. And you’ve been coding on science. India. And of course we always bring in the international experts so that India can learn and also this is a good opportunity for India to expand their capacity to outside countries. Thank you.

Thanks, Miyu. I actually had a follow -up question for you that would have touched on this de -risking and catalyzing investment topic. Maybe I’ll let you ask you to repeat that now, but let me just add our digital sector office being fairly new. They are getting a lot of demand for guess what? Data centers. And, you know, we welcome that. We have conversations with government but I’m truly impressed with the conversations at the summit here. Earlier this morning I attended one where the state of Telangana was sharing what they’re thinking about and they’re really quite cognizant of the kids to school not many kids fit into a Ferrari milk doesn’t make sense so we need to look at what type of compute is needed for what and I think we in ADB are also learning more and more how to engage in these conversations properly we’re learning alongside everybody else in this room probably and that’s an important distinction to make because it will influence the financing bit how much do we need, what do we actually need and when and how do we make that investment sustainable just wanted to add that it’s an insight from this morning that I couldn’t not retail.

Let me go back to Indonesia and ask you about cutting -edge skills because you’re in that space. I found it very interesting that you’re actually, you said co -president or co -chairing this platform where you bring together private sector, education sector, government. And as you are looking at that, how is your organization doing it in practice? How do you bring these people together and get them into action mode? How do you do that?

Okay, thank you. Very important question here, I think. So I would like to say in three pillars that what we are doing, especially that we chair the AI ecosystem in Indonesia where the government, the industry are involved in the AI ecosystem. Within an academia as well as the… civil society as well as media we call this pentahelix platform we discuss about I think three pillars first one is the talent certainly second one is infrastructures and the third one is how basically we can articulate use cases towards all the services public services and businesses as well so Indonesia for talent we have set our target quite ambitious that we want to have at least 12 million talent by 2030 and for us this is something that are uh fairly challenging, considering that we are still lacking around 3 to 5 million talent as of now, right?

So what we are trying to achieve together with the whole ecosystem is to establish an academy, the Korika Academy, where we promote to not only upskilling and reskilling some of the civil servants and other workers, but we are also looking at how we can train the trainers. We work with several of our friends. I will note here that Elevate Indonesia, for example, part of the… Microsoft and many others big tech that are there works together with our ministry to establish this program for Thailand

and it’s a digital academy or is it a physical?

it’s a digital academy with the LMS learning management system and many other things we also established the Kodika chat actually it’s a chatbot for this training and upskilling program that we do with the government beyond Thailand basically we are aggressively looking at how we can nurture this talent to work in data centers, in many others startups and incubators as well as to establish some of the most diverse demanding use cases So the third one is we try to work on climate health nexus in establishing how we counter and predict the climate sensitive infectious disease such as malaria and dengue. And we have established for the past three years the Climate Smart Indonesia which have attracted many of the universities as well as NASA pollution and air quality programs to look into these use cases.

So we can basically reach out to many of the areas where the… …the health, the disaster prone area because Indonesia is a supermarket for disaster. You can have the hydromelectorological disaster, you can have ecological, you can have many things. So you need to…

I’m not buying any of them.

Of course, we don’t want to be shopping.

So really amazing this focus also on the use cases, right? And prioritizing those that match with your country needs. Yes, thank you. Give the highest impact, right? Super. I’ll turn back to Uzbekistan and just wanted to ask you to elaborate a little bit in terms of private sector capital mobilization. You have all these ambitions you shared across the board really in terms of infra, in terms of skills and so on and so forth. Uzbekistan as an economy has still… a good chunk of traditional economy but also has a very active startup sector that I’m learning more and more about how dynamic people are around the region, Central West Asia going and finding scalable solutions but these are the still growing companies for mobilizing capital for your infra you’re going to need the big ones or you’re not going to need the international partners or what are you thinking about this private capital mobilization, what’s your strategy there?

First of all I should mention that according to the documents adopted by year 2030 we are planning to attract around 1 billion USD for investments for creating AI related digital infrastructure and part of this goes to creating data centers and we’re going to need to And also we’re working with our Chinese partners also. It’s the biggest IT company, Huawei. So they’re also involved in creating AI ecosystem in Uzbekistan. Mainly, first one is upskilling public servants to help them to adopt AI adoption and also creating the necessary training programs for the specialists and also creating the AI infrastructure like data centers, data lakes. And also we need to get, we are transferring to 5 .5G and also working on 6G also with Huawei.

So, yes. And also, as you mentioned already about startups, we are developing our own. startup ecosystem and we established many venture funds funds of funds and also there are many emerging private funds so they are now trying to invest in startups attracting private funds, private investments so currently we have allocated around 50 million USD for AI startups so they are already providing services both for public and for businesses so trying to balance and attract all the stakeholders of the ecosystem

Fantastic, so you’re mixing also your public funds that you invest for example in the fund of funds and then bringing in more investment domestically from your investors but also from abroad That’s amazing. And then having large industry partners that are interested in the market, bringing them in like Indonesia did with some of the hyperscalers. You are bringing in Huawei and Chinese partners. So basically it’s a mix of different strategies you mentioned. Also, that’s fascinating. Again, Uzbekistan being one of the larger countries in Central West Asia and Indonesia, both fairly large in their region. And then, of course, again, I want to come back to this point about diversity of country context. That’s both a challenge but also an opportunity.

I mean, for us at ADB, it adds, of course, complexity because we need to respond to these different needs. But from the perspective of WTO, is there like a specific area such as maybe interoperability standards or AI talent mobility? Or the shared data set? joint research, where do you see regional cooperation making the biggest difference? I

think that it’s a bit of a matter of context, right? At the regional level and at the national level. We’ve talked about the divide in the digital divide and how do we overcome that and the role of infrastructure and skills and the rest. And at the WTO Secretariat, we’ve been very concerned on this issue. And so we partnered with the World Bank and we did a study called Digital Trade in Africa, a general one, and then we did some country pilot studies to look at the situation. And we did see that some of the regional work, like the ACFDA and the digital protocol, really made a difference in how it helped bring them along and to set a certain standard in many of the countries that we studied.

Then we did a similar study with the World Bank in Latin America and the U .S. The Caribbean and the Inter -American Development Bank partnered with us. And we saw there that the situation was a bit different, more diversity in terms of regulation and trade policy, infrastructure needs. So there’s basically not one size fits all. But we have seen regional banks playing a very important role in helping countries that want to go in the regional way. I know ASEAN has done important work in AI policy, for example, and other regions are also working in that sense. And I do think that that brings economies of scale to a certain extent. It helps you resolve questions on electricity sometimes.

And so I think there’s a lot of opportunity and further work to be done at the regional level.

Thank you. And I think with the regional cooperation integration agenda being also top of mind for ADB, I just ask my colleague also to… tell me a little bit about her perspective. Of course, she represents ADB in a very large economy in South Asia, but we do have regional cooperation happening around the region. Mio, what do you see as opportunities with regards to regional cooperation integration on this digital infrastructure space?

Right, thank you. So again, ADB, we support India. I’m in India. Our office covers India. So we are here to support Vixie Bharat so that India can grow at the pace to become a developed country by 2047, and the AI is a necessary means to do that. But again, as everybody knows, we are the regional bank. Nobody around us should be left alone or left behind. So ADB, through this kind of forum, has to be a catalyst. A catalyst for the global south. So we are here. Of course, there are many countries who cannot invest in scale. What are the solutions? So we are here to support the solutions, and also we support big tigers like India to support those countries too.

That’s number one. And number two is the balance approach. When we talk about regional cooperation or the work in a small country, I was quite shocked about five years ago. I went to the small neighbor country here, and I was working in the agriculture sector, and I was proudly introducing, I want to introduce aquaculture using the AI -based fish feeding system. And my negotiation ended in three seconds because the government said, no, we are interested in employment. What are you talking about? What AI -based feeders will just reduce the people who are going to work there? So that is a big lesson learned for me. We need an ecosystem, but even we talk about AI, the solution may be elsewhere.

so as you introduced the skill is super important and since that understanding again going back to India we’ve invested more than like 5 billion in the skill including the PM set and working over 10 states and now AI based skill is the big part of it so we are always mindful that the regional cooperation and we should not forget should not leave any country to be left behind but solution again may not be as direct as we expect thank you

thank you Mio and we have one minute left on the clock that throws a spanner into my closing with the thought that AI may not be the solution for everything but I think it’s a fair ending looking for a name We need to understand the problems and see how AI, if it can be deployed, if it can make a difference, how it should be supported through skills development, infrastructure investments, regulation. So I want to thank my panel for a very interesting tour de force of this topic. Also thought I’d take the opportunity to thank the audience and India for hosting this amazing summit. As ADB, we’ve been proud to be a partner of it, and it’s been truly fascinating, and we’re quite proud to have been part of this journey.

Thank you all for attending, and thanks to the panel. Let’s give them a round of applause for sharing their views. Thank you. Thank you. Thank you very much. Thank you. Thank you.

that the race towards artificial intelligence is no longer a theoretical pursuit. As billions and maybe trillions now of dollars are getting deployed to push the frontier of artificial intelligence, the technology is now advancing rapidly. And safety is not keeping pace with it. There are wonderful opportunities on the other side of this quest. There are also big risks. And so that’s the purpose, that’s the reason AI Safety Connect was founded. AI Safety Connect is there to help shape the frontier AI safety and secure agenda towards what I would frame as commonsensical AI risk management. AI Safety Connect has been founded to encourage global majority engagement into frontier AI safety. And AI Safety Connect, has been connected to showcase Concrete’s governance coordination mechanisms, tools, and solutions.

So how we do this? We convene at each AI summit. So last year we started in Paris, this year in India, next year we’re going to be in Switzerland. But we also convene at the UN General Assembly, right? We need a faster tempo for these safety discussions, so every six months we have this global convening. We also do capacity building, and we also do trust building exercises at times behind closed doors. Well, this week in New Delhi has been an intense one, an impactful one. On Tuesday we had a full day of panels, conference, solution demonstrations, and closed -door workshop discussions on some specific nuts to crack to advance AI safety. We, for example, at the privilege of, hosting Prime Minister Dick Schuh from the Netherlands on stage to deliver a special address on the role of top leadership in advancing AI safety.

We also engage with industry, engage with academia. of India and abroad. So we’re an extremely busy week beside our main event. We had this closed -door discussion that I was mentioning yesterday and today, this closed -door scientific dialogues. We’re going to publish the results soon that brought together senior industry leaders to discuss shared responsibility for AI safety. Well, obviously, none of this would happen without partnership. And we want to thank our co -hosts, the International Association for Safe and Ethical AI and its director, Professor Stuart Russell, to whom I will hand over the floor in a few minutes, and the Digital Empowerment Foundation who is anchoring us at the grassroots here with Osama Manzar,. We’ll close the session later on.

And we obviously want to thank our sponsors and supporters, starting with Sympathico Ventures. Eileen Donahoe will moderate that panel and we’re thankful for that. The Future of Life Institute, Ima and Yann, who’s been supporting this effort, and the Mindero Foundation, whose team is here as well with team. And it’s great to have your support and we are thankful for that. So today we’re about to hear from His Excellency Matthias Korman who’s the Secretary General of the OECD We’re going to hear from Her Excellency Minister Josephine Theo who’s the Minister for Digital Development and Information at the Government of Singapore. Thank you for your continuous support, really appreciate that Same for Jann Tallinn who’s the AI investor but also a founding engineer at Skype and the co -founder of the Future of Life Institute And last but not least, we also have Minister Teo who’s going to be with us from Malaysia Minister for Digital Development and Information Thank you Minister as well as Vice President Kim for Digital and AI at the World Bank So an extremely important conversation to have And before we welcome you to the stage I would like to hand over the floor to Professor Stuart Russell to say a few words and to speak about also what’s happening next week in Paris Thank you so much.

Thank you very much, Cyrus and Nico. So as Nico mentioned, the International Association for Safe and Ethical AI, or ICI, the world’s worst acronym, is a global, democratic, scientific and professional society. We have several thousand members and approaching 200 affiliate organizations. Our mission is to ensure that AI systems operate safely and ethically for the benefit of humanity. And as Nico mentioned, our second annual conference will take place in Paris starting on Tuesday. It’s still, I think, possible to register, but we’re already up over 1 ,300 people coming. It’s at UNESCO headquarters in Paris. Thank you. So achieving this mission of ensuring… that AI systems operate safely and ethically is partly a technical challenge. How do we even build systems that have that property?

But also a governance challenge. How do we ensure that those are the systems and only those systems get built? And this panel is mainly about this second challenge. And I think it’s one on which global coordination is essential because the harms, whether it’s psychological damage to the next generation or loss of human control altogether, those harms cross borders. And we must coordinate to make sure that they don’t happen or they don’t originate anywhere. And it’s, I think, fitting that we are having this summit here in India, which has really, among other things, championed the idea that everyone on Earth should have a say. And so with that, I will hand over to Eileen. Thank you very much.

Thank you, Stuart. So Dr. Eileen Donahoe is the founder and managing partner of Sympathico Ventures. She’s also the former U .S. Special Envoy and Coordinator for Digital Freedom and Ambassador to the UNHCR. Eileen? Welcome the speaker on the floor. Please, Your Excellency, Mr. Mattias Korman, Mr. Gobind Singh Deo, Mr. Josephine Teo, and Mr. Jann Tallinn, as well as Mr. Sangbu Kim, join us on stage.

Okay. Given this remarkable panel and the very short time we have, let me very briefly frame our discussion and get right to our speakers. So we’re here to share. Views on the opportunity for policymakers to impact international AI governance. As the race towards AGI and superintelligence intensifies, AI safety advocates face a compounding challenge. The technology is advancing rapidly and being deployed with minimal guardrails, while the risk management processes that do exist are either ill -adapted to the magnitude of the risk, fragmented across jurisdictions, or insufficiently binding on developers, deployers, investors, and regulators. The result is an unharmonized governance landscape that fails to shape the behavioral incentives. Of those building and funding frontier AI. Economies, governments, and societies do not respond well to such mixed signals.

While much of the discourse on frontier AI safety has focused on AI superpowers, there’s an urgent need for deeper international diplomacy on the most… extreme risks. At this juncture, middle powers and global majority states can’t be seen as peripheral actors in this landscape. Through pooled resources, market leverage, normative influence, and regulatory innovation, they can shape the direction of global AI practices and safeties. Leading from the middle may turn out to be a more powerful approach than previously anticipated. Whether or not that collective power is exercised now will determine whether international AI governance moves from the rhetorical level to the real -world impact on safety. This panel will aim to identify present -day coordination gaps in the global AI practice and the global market.

We will also look at the role of global AI in international AI safety and highlight practical steps policymakers can take in the coming months to close them. So to our panel, I’ll start with Secretary General Corman. The OECD has done remarkable work over the past decade, developing consensus on the OECD principles, providing a definition of AI systems that has resonated internationally, and playing an international role in operationalizing the Hiroshima International Code of Conduct. Along with those foundations, we now have the International AI Safety Report and the Singapore Consensus on Global AI Safety Research Priorities. With these principles, definitions, and frameworks in mind, two -part question for you. First, what are the key lessons learned from the process of building consensus and then implementing these frameworks?

And then second, looking ahead, what’s the most critical? What’s the most critical piece of coordinated frontier AI safety infrastructure we should be building now? Some have called for an international incident response center, and we’re all curious whether you think that should be a priority and achievable. Just some small, easy questions.

In terms of what is the key to success, what is the most important lesson on looking back on what we need, trust is built through inclusion and on the basis of objective evidence. And, you know, I think what we’ve learned over the last few years is that bringing together all the relevant actors, governments, companies, civil society, technical experts, is what we need to do. I mean, each has a different perspective and different imperatives. I mean, markets reward the private sector for speed, scale, and innovation. While governments must manage risk and protect the public interest without stifling progress. But a challenge, and it’s been mentioned in some of the opening remarks, a challenge for policymakers in this context is that AI is moving much faster than policy cycles have traditionally moved, which easily then creates gaps between innovation and progress and opportunity, but necessary oversight, mitigation and management of risk.

But all sides in this conversation do share an essential common interest, and that is to ensure that the systems that are developing are trustworthy, because without public trust in the end, even the most powerful AI tools will struggle to gain broad adoption. So that means that occasionally, and it’s not always popular with everyone, but occasionally we should slow down. Occasionally we should actually pause. Pause, test, monitor, audit, share information, and take the time and invest in building confidence that these systems can work as intended and respect fundamental rights. So that’s sort of, I guess, the first point. another critical lesson involves international consistency and this is part of the reason why these sorts of summits are so important is to really facilitate these conversations among countries and among different jurisdictions because national priorities can vary quite widely and there’s of course fragmentation and compliance cost related risks and at the OECD really what we’ve been doing for six decades now across different policy areas is to try and reduce fragmentation and by achieving alignment around key principles, building shared evidence and facilitating the necessary conversations to develop a more coherent better coordinated approach moving forward and on AI I mean we’ve developed the OECD principles which were first adopted in 2019 and which are now adhered to by 50 countries around the world and that was really the first globally recognized baseline for trustworthy AI The OECD’s lifecycle definition of an I .I.

system has since shaped policy frameworks from the EU I .I. Act to U .S. executive orders. And we’ve had just earlier the meeting of the Global Partnership on I .I. co -chaired by Korea and Singapore. We’ve got the OECD I .I. Policy Observatory, which is sort of essentially the broad gamut of all of the different policy approaches around the world to provide countries and industries with data and evidence on what’s being done, facilitating peer learning, and trying to take some of the politics and the rhetoric out of it, but really looking at the facts. Now, looking ahead, and you sort of ask a question here about what to do about the risk. I mean, the most critical piece of frontier I .I.

safety infrastructure is coordinated. transparency and incident reporting. I mean, the Hiroshima I .I. Process Code of Conduct and its reporting framework launched at the I .I.’s Action Summit in Paris last year. You know, that’s a promising step, and we’ve got to continue to develop that. Since their publication, 25 organizations across nine countries have already submitted detailed reports on how they manage I .I. risks, offering for the first time a comparable view of developer practices across jurisdictions. The next stage is to strengthen information sharing on I .I. failures and near misses. The GPI I .I. Common Framework for Incident Reporting aims to help us collectively learn from mistakes before they scale globally, and over time, this could evolve into an international I .I.

Incident Response Center, coordinating alerts between governments and labs without exposing companies to commercial or legal penalties for reporting in good flight. Finally, we do need to scale access to practical safety tools. With global partners, the OECD recently launched an open call for open source safety and evaluation tools hosted in the OECD .ai catalog of tools and metrics to make a trustworthy AI easier to implement in practice. I mean, these are some initiatives to form the foundation of a more transparent, data -driven, and interoperable AI governance ecosystem, and

Excellent. Minister Teo, a number of questions for you, but let me start with the fact that Singapore occupies a very distinctive position in the global geostrategic landscape as a pro -innovation, advanced knowledge economy, with deep commercial and diplomatic ties to both the U .S. and China. Thank you. As the race to AGI intensifies and bilateral tensions mount, is there a role for Singapore and other middle powers to play in bridging the coordination gap to keep scientific and safety channels open? And also, what’s the most important step middle powers can take in the next 12 months to help establish a shared minimum understanding of frontier safety?

Well, thank you very much for that question. I think there is no running away from the fact that for smaller states, and that includes Singapore, the technology that our companies, our citizens are going to rely on do not originate from our shores. So they don’t necessarily come within our jurisdictions. We don’t always get to set the rules. Having said that, I do believe that we’re not without. Thank you. agency. It doesn’t mean that we take a step back and just let things happen to us. There are still things that we can do. One of the most important things I think as policymakers is for us to think about what it takes to translate what we know from science into policy.

And I wanted to just say why this is so important. In our case, as policymakers, the key questions will always be, are the policies that we make effective? And also, policies always come with trade -offs. With the question of effectiveness, there is always a need to understand what actually works, as opposed to what looks good on paper. With the question of trade -offs, it’s about understanding what we lose as a result of whatever safety aspects it is that we choose to put in place. And whether we can minimize them, can we mitigate them? Now, in areas where safety is the objective, we can’t just go with gut. We can’t just go with speculation. You take, for example, in my previous life, I was working on promoting Singapore’s Air Hub.

And we had to deal with a question of aviation safety. We were expanding our airport. It was going to carry many more passengers in and out of the country. But we are limited by number of runways. And in landscape Singapore, you can’t just click your finger and say, let’s build a new one. It’s a long runway. It’s very expensive anyway. Then there is the question of what do you do when you have these jumbo jets like A380s? Because each time an A380 hits the runway. It creates so much of a blast that you really need to create more distance between the A380 taking off and the next aircraft that is scheduled to take off.

Now, this is not a question that the transport minister can just decide on a whim. The air traffic management has to decide on its policy of how much distance is considered safe between landings or rather between takeoffs. And to answer this question, you really need to invest in the research. You need to invest in understanding the tests. So the science is one part of it. But between science to policy, you are actually going to need a lot of time. You are going to need a lot of tests. You are going to need a lot of simulations. you need to understand whether the distances that you decide are safe works well in a thunderstorm, a tropical thunderstorm.

Does it work just as well in a snowstorm? Well, we don’t have snow in Singapore. But you think about the airline that operates this. If each country that they fly into has a different safety distance, that creates some difficulty. So we therefore think that not only is there a need to invest in understanding the science, not only is there a need in understanding what testing looks like, what good testing looks like, there is also a need for us to think about what standards that will eventually be interoperable, what do they look like, which is why we think that international efforts, the collaboration that… that is being carried forward by the OECD through the Global Partnership on AI, the AI Safety Connect effort, and also ICI.

Where is Stuart now? Those kinds of efforts, you can’t do away without. At the outset, there is likely to be a bit of a fragmentation. And the trade -off with not having these conversations is that we are not even going to make advances in AI safety. And I don’t think that that’s a very good place for us to be in. It doesn’t give us the assurance that we can deliver to our citizens. And it does not create a foundation of trust that will eventually help us to push ahead with the use of this technology on a wider scale. So that’s how we are thinking about it, Aileen. Thank you.

So let me turn to Minister Gobind from Malaysia. and note that under your leadership and Malaysia’s 2025 ASEAN chairmanship, Malaysia succeeded in placing AI at the center of ASEAN’s agenda by establishing the ASEAN AI Safety Network. Malaysia is now finalizing its own AI National Action Plan, and Malaysia’s AI Governance Bill is expected in Parliament in 2026. So this dual -track approach of building national capacity while leading regional coordination represents a model of middle power agency that other countries are watching closely. So what lessons do you think other middle powers can draw from Malaysia’s experience? And on the ASEAN AI Safety Network, we have to note that operationalize and it will require sustained political will. technical capacity and resources.

So what concrete steps must ASEAN take in the next 12 to 18 months to ensure that this isn’t just aspirational?

Online fraud, for example, scams, you have deepfakes today, you have huge concerns about certain vulnerable groups that are going to be impacted, children, older folk and so on and so forth. So this is something that stretches across the region. How do we deal with it in a coordinated way and ensure that the conversation doesn’t just stop with the government of the day, but it’s a conversation that expands over a period of time with clear policies that we can actually execute. The second layer that I think we need to think about is in the event there’s a need for execution. When we speak about risks in AI and we speak about how we’re going to govern these risks, we often talk about standards.

We often talk about regulation. We even speak about legislation at times for areas that pose higher risks. But ultimately, it really comes back down to you making sure you have an agency that can enforce it, because you can have the best standards. regulations and legislation but if there is no institution that’s really able to implement those standards to ensure that they are properly implemented and also to ensure that rules for failure to implement are enforced then those standards regulation and policies are really going to be just strong on paper but they’re not going to really have that impact that you need. So again, how do you build this mechanism across ASEAN where every country strengthens themselves domestically first and then moves across to the ASEAN member states and hopes to learn from their experiences so that we can together move ahead in this new world of AI and I think the threats that we anticipate in future.

Now the third part which is really important is also ensuring that whilst this goes on, you create those policies, you have institutions that enforce and the discussions persist at an ASEAN level. I think what is important is also to have that expertise looking at what comes next. We must make sure that our countries are prepared for the risks that are to come with the next generation technology. This is important because you don’t want a situation where new technology is adopted and there are risks that come with this new technology, you’re not prepared. I think that’s something we want to avoid and that’s the reason why I come back to where I started off. We really need to look at building institutions that have the expertise and of course are able to sustain as we go along and to build and deliver something that’s impactful.

Sorry, but that’s in short what we’re doing in Malaysia today.

Excellent. Thank you so much. Okay. Let me turn to Vice President Kim and talk about the World Bank, which has been at the forefront of digital public infrastructure, helping countries leapfrog legacy systems. We note that frontier AI systems, though, are arriving in the global south under very different conditions from previous waves of technology and governments are under pressure to deploy AI systems quickly. often using models that haven’t been adequately tested, let alone certified for their context, languages, or risk tolerances. So how can the World Bank help Global South countries move from being passive recipients of frontier AI to active shapers of safety and reliability requirements before the systems are deployed at scale?

Thank you. In one word, definitely we need to make our clients well prepared from the scratch. When they design the AI systems, definitely they need to design the safety architecture within the system. That’s very, in general, that’s very correct. But real challenge is that… nobody can really expect a new type of new threat especially our some countries in a low capacity it is really hard to figure out what that will be so that’s the in order to tackle that type of irony and dilemma we need to very closely working with very developed economies company and government and very high end examples so that we can really well connect those good examples to the developing world so one partnership is one of the good examples we are helping our country for example some big tech company who is running some red teams so that you they are trying very hard to attack their system in advance by fully utilizing AI.

So through that type of practice and experiment, they can learn how to prevent the AI attack in the future, which is pretty much possible. So in this way, it is inevitable for our developing countries to keep track on the new trend and new innovation, even in this safety protection area. It is the only way. So I have to admit that this constraint. But think about this. Some anecdotal story in East Asia, in China and in Korea, there’s two models. Merchant who is selling two products. Number one is. sphere. And then they keep saying that this sphere is so strong so it can get through any kind of shield. So this is one vendor. The other vendor is selling shield.

And then they are saying that this shield is one of the most safe and strong shields. No sphere can get through this shield. This is exactly an ironical situation. If you think about AI, AI attack is the sphere. AI is so strong and smart and really capable so it can get through and hack any system with high -end intelligence and knowledge. But good news is that on the other hand, we also can build strong protective systems. by fully utilizing AI. So this is one good news, but the constraint is that we do not clearly know how AI can really evolve to fully protect those big attacks in the future. So in order to solve this type of ironical situation from the developing world point of view and from the World Bank point of view, this is the only way to very closely work and collaborate and learn from the advanced technology and advanced company and advanced country.

Thank you so much. Last but not least, Mr. Jan Tallinn, you occupy a very rare position in this landscape as a founding engineer of Skype, an early investor in DeepMind and Anthropic, and you’re also the co -founder of the Future of Life, which last October released a statement on superintelligence. calling for prohibition on superintelligent development until two conditions are met. Number one, broad scientific consensus that it can be done safely and controllably, and second, strong public buy -in. Let’s just ask the hard question. What would an effective prohibition look like in practice? How could that work?

Thank you very much. So I think I’m kind of like a little bit different from the people on this panel. And that too, I guess. That I’m kind of, my main kind of threat vector about, my main worries about future are less about like how AI is being deployed and diffused and taken into practice. And I’m way more worried about what is happening in the labs, in the top AI companies. I’m not sure what the future is going to look like. because they are now in a cutthroat race to build something that is smarter than they are. They are in a cutthroat race to build superintelligence. And, like, I mean, we just saw yesterday the picture where, with a photo of it, Narendra Modi, Dario Amadei, and Sam Altman refused to link hands.

I mean, this is, like, indicative. We also saw both Dario and Demis Hassabis call for a slowdown in Davos last month. They just can’t do it alone. And I think there are, like, two reasons why it’s, like, an unfortunate situation. One is that the U .S. as a country is conflicted. They basically rely on AI for their economic and competitive power. So they are, like, very hesitant to, kind of, meddle with now. cutthroat situation in AI companies and the rest of the world really doesn’t understand how big danger they are now. So it’s part of the reason why we did the superintelligence statement is to create awareness that there is increasing political demand to do something about this situation.

We now have more than 130 ,000 signatures which is like many times more than we had done our original six months post letter had in 2023. So yeah, that’s like if there was enough pressure, I think clearly like the rest of the world is still kind of more powerful than the kind of leading AI countries. There are more people, there’s more economic power, etc. So if there was like enough pressure this could be solved. Like the way I put it is that it’s super hard to do like a $10 billion project. it’s impossible to do it if it’s illegal. So having these trillions flow into AI actually makes it easier to govern than harder.

So I’m tempted to follow up with a question about investors and their potential role in this. They are obviously playing a decisive role in shaping the incentives, but they’re largely absent from the governance conversation. So what would it take to bring investors meaningfully into the safety conversation?

So, yeah, I think the answer is kind of simple. I don’t think investors play much of a role anymore because the leading AI companies now are kind of above the level where private investors can influence them. They will now IPO soon. And if you are like an IPO market, there is… like, like, so level playing field, which means that like, if somebody’s not funding, somebody else will. So I don’t think investors, investors could have affected things, but like, five, 10 years ago.

Great. Okay, so since we’re running short on time, I’m going to ask one question, and ask you all to answer it, which is about the 12 month window. Oh, the very shortly, each shortly. Many in the AI safety community believe we have a narrow window, perhaps 12 to 24 months before frontier AI capabilities advance beyond our ability to evaluate and govern them. So what would you recommend is prioritized between now and we’re basically in the next year to two years, each of you to enhance safety? and security?

I think there are two, really. I think the AI safety research priorities need to be refreshed because the field has moved so quickly. The Singapore consensus identified a set, but as soon as they are published, we recognize that they will be out of date. So we need to refresh it. That’s why we’re going to have the second edition, you know, worked on. Hopefully in a few months. The second thing I think is that we can’t just keep thinking about frameworks, you know, and guidelines. At some point, we need to be able to introduce better testing tools. And until we are able to do so, the companies that are developing and deploying AI models, they also don’t have a very practical way of giving assurance.

So I’d like to see in the next 12 months some further advancements. In those two areas.

I’ll be really quick I know there’s always a temptation in these sorts of conversations, what is the one thing that can sort of fix it all and the truth is there’s not one thing we’ve got to go as fast as we can to play catch up to a degree but we’ve also got to go as comprehensive and as deep as we can there’s just no alternative, there’s catch up to be played, we’ve got to put a real effort and it’s got to be right across the board and I don’t think that you can just say there’s the one thing that will make us all safe and it’s going to be okay.

Minister Gobind?

I think as I said earlier, we need to start thinking how we can build structures and perhaps institutionalize this entire conversation about building security around AI and its governance in this regard, we have to understand that things are going to move very quickly and you’re going to see new technology develop very fast which brings new risks as well, so in that regard, you’ve got to build something that’s sustainable and I think in order to do that, institutionalizing it should be a priority.

everyone is really rushing for ai system development ai solution development that means ai is currently ai safety measures currently under invested so i really like to urge all of us to think about this is not free you know things we need to spend some money to protect the system in advance from the scratch when you design the system so that means we should allocate some money to fully invest in in the

Jann Tallinn?

so slow down we really need to slow down that the companies are asking for it and if we like instrumental to that would be basically transparency like more people should know what the leaders of ai companies know in order to basically understand how crucial the slowdown now is

okay great well i believe we have a little bit of a close coming and thank you all so much i wish we had had a day to talk about all of these issues. But thank you so much. Thank you very much.

Thank you very much, Eileen, and this fantastic panel, excellencies, colleagues, friends. What we’ve heard today confirms something important. The coordination gap frontier in AI safety is real, and it is urgent. And as we’ve discussed today, it is closable. And before I hand over the floor to Osama Manzara to close off for a few minutes of remarks and reflection, I’d like to invite you all to the United Nations General Assembly next edition in New York, where we hope to organize the fourth edition of AI Safety Connect, and hopefully with many of the great policymakers and leaders we have heard from today, to carry forward that collective effort. Osama, the floor is yours.

Well, thank you very much. And we are one of those absentee co -organizer in this one. So, you know, because being a local, but I just want to I mean, apart from thanking each one of you who didn’t get up and, you know, go out of the room. And every one of you who gave all the safety remarks before usage of AI on behalf of 40 million people that we have reached out in the last 23 years. And billions of the other people whom we are going to work for. I want to suggest that the entire safety aspect of AI should be more from please save people from AI. Right. Because that’s the safety like it’s a car on the road.

You know, we have to save people before you teach people how to think. So we also have to keep a very, very strong thing. How do we save human intelligence from artificial intelligence? And how do we inbuilt in the safety guards and all the ethics and all the all the, you know, policy playbooks? Thank you very much. Thank you. Bye. Thank you. Thank you.