What policy levers can bridge the AI divide?

Summary

This panel discussion, moderated by Ebtesam Almazrouei, focused on bridging the AI divide across different nations and regions, featuring government officials and technology leaders from Zimbabwe, Costa Rica, Lithuania, the Philippines, and Smart Africa. The central theme emphasized that addressing the AI divide requires first tackling existing digital infrastructure gaps, as AI systems depend on robust broadband connectivity and computing power.

Lacina Kone from Smart Africa highlighted that 2.6 billion people remain unconnected globally, and stressed the importance of developing AI solutions in local languages, noting Africa’s 2,000+ languages and the need for culturally relevant large language models. Zimbabwe’s Minister Tatenda Mavetera outlined their “PSL-I²” framework focusing on participation, privacy, skills, leadership, legislation, and local solutions, with particular emphasis on agricultural applications using AI for precision farming and weather monitoring.

Costa Rica’s Vice Minister Hubert Vargas Picado emphasized equity as fundamental to bridging the AI divide, describing their national AI strategy developed with 50+ institutions and programs targeting rural and indigenous communities. Lithuania’s Jurate Soviene discussed the challenge of digital refugees in aging European societies and proposed regulatory sandboxes for AI innovation, similar to their successful FinTech approach.

The Philippines’ Emmy Lou Versoza Delfin addressed the unique challenges of implementing AI across 7,000+ islands, highlighting their National Broadband Program and free Wi-Fi initiatives reaching 17,000 sites. Common themes across all presentations included the need for international cooperation, inclusive policies, skills development, and ensuring no one is left behind in the AI revolution. The panelists agreed that success within one year would involve more localized AI solutions, improved multilingual capabilities, reduced biases, and greater accessibility for underserved populations worldwide.

Keypoints

## Major Discussion Points:

– **Foundational Infrastructure Challenges**: The discussion emphasized that bridging the AI divide requires first addressing basic digital infrastructure gaps, with 2.6 billion people still unconnected globally. Panelists highlighted the need for broadband connectivity, electricity access, and meaningful internet usage before AI implementation can be effective.

– **National AI Strategies and Frameworks**: Each country representative shared their specific approaches to AI development, including Zimbabwe’s “PSL and I²” framework (Participation, Privacy, Skills, Leadership, Legislation, and Local solutions), Costa Rica’s inclusive strategy developed with 50+ institutions, and the Philippines’ National AI Innovation Strategy (NAIS).

– **Skills Development and Digital Literacy**: A major focus was on addressing the skills gap through upskilling, reskilling, and creating inclusive programs for different population segments, including “digital refugees” (those who find digital tools intimidating), rural communities, and underserved populations.

– **International Cooperation and Localization**: Panelists discussed the importance of collaborative approaches while emphasizing the need for localized solutions, including developing AI in local languages (Africa has over 2,000 languages) and creating region-specific applications rather than adopting one-size-fits-all solutions.

– **Sector-Specific Applications**: The discussion highlighted practical AI applications, particularly in agriculture (precision farming, weather monitoring, crop analysis) and e-government services, as key opportunities for developing nations to leverage AI for economic development.

## Overall Purpose:

The discussion aimed to explore strategies for bridging the global AI divide by sharing best practices, challenges, and collaborative approaches among different nations and regions. The goal was to identify practical mechanisms for ensuring equitable access to AI benefits across developed and developing countries.

## Overall Tone:

The discussion maintained a collaborative and optimistic tone throughout, with participants sharing experiences constructively and focusing on solutions rather than dwelling on challenges. The tone was professional yet inclusive, with speakers acknowledging difficulties while emphasizing opportunities and the urgency of action. There was a consistent theme of “leaving no one behind” and a shared commitment to international cooperation, creating an atmosphere of mutual support and shared responsibility.

Speakers

**Speakers from the provided list:**

– **LJ Rich**: Moderator/Host (introduced the panel at the beginning)

– **Ebtesam Almazrouei**: Executive Director of the Office of AI and Advanced Technology at the Department of Finance, CEO and Founder of AIE3, Chairperson of the UN AI for Good Impact Initiative (Panel Moderator)

– **Lacina Kone**: Director General and Chief Executive Officer, Smart Africa

– **Tatenda Annastacia Mavetera**: H.E. Dr., Minister, Ministry of Information, Communication, Technology, Postal and Courier Services, Zimbabwe

– **Hubert Vargas Picado**: H.E. Mr., Vice Minister, Ministerio de Ciencia, Innovación, Tecnología y Telecomunicaciones, Costa Rica

– **Jurate Soviene**: H.E., Chair of the Council, the Communications Regulatory Authority of the Republic of Lithuania

– **Emmy Lou Versoza Delfin**: H.E., Director, Department of Information and Communications Technology, Philippines

**Additional speakers:**

– **H.E. Mr. Solly Malatsi**: Role/title not clearly specified in the transcript, but appears to be mentioned in context of collaboration efforts

# Bridging the AI Divide: A Global Perspective on Equitable AI Development

## Executive Summary

This panel discussion brought together government officials and technology leaders from across the globe to address bridging the AI divide. The panel featured speakers including Lacina Kone, Director General of Smart Africa; H.E. Dr. Tatenda Annastacia Mavetera, Minister from Zimbabwe; H.E. Mr. Hubert Vargas Picado, Vice Minister from Costa Rica; H.E. Jurate Soviene, Chair of the Council from Lithuania; and H.E. Emmy Lou Versoza Delfin, Director from the Philippines.

*Note: This summary is based on a transcript with significant audio quality issues, resulting in some unclear or fragmented speaker responses. Where content was unclear, this has been noted accordingly.*

## Defining the AI Divide: Country Perspectives

The discussion began with each speaker defining what “bridging the AI divide” means from their regional or national perspective, revealing diverse priorities based on different development contexts.

**Smart Africa’s Continental View**: Lacina Kone provided perhaps the clearest framework, emphasizing that “AI sits on broadband” infrastructure and highlighting that 2.6 billion people remain unconnected globally. Kone distinguished between the coverage gap (15-20% of the connectivity challenge) and the usage gap, which represents the larger barrier. Even where broadband infrastructure exists, people aren’t utilizing it due to affordability issues, lack of smart devices, absence of local content, and inadequate digital literacy.

**Zimbabwe’s National Strategy**: Minister Mavetera outlined Zimbabwe’s approach, mentioning what appears to be a framework focused on ensuring “no one left behind.” The country appears to prioritize agricultural applications and rural connectivity, though specific details were unclear in the transcript.

**Costa Rica’s Regional Leadership**: Vice Minister Vargas Picado described Costa Rica’s inclusive national AI strategy, developed with multiple institutions and emphasizing equity and inclusion. He mentioned their role in developing tools to assist Central American and Caribbean neighbors, though technical details were fragmented in the transcript.

**Lithuania’s European Perspective**: Jurate Soviene introduced a framework for understanding digital inclusion challenges, categorizing populations into digital natives, immigrants, and refugees. Digital refugees—those who find digital tools intimidating—represent a particular challenge in aging European societies.

**Philippines’ Archipelagic Challenges**: Director Delfin highlighted the unique challenges of implementing AI strategy across 7,000+ islands, requiring particularly inclusive approaches. She mentioned their National AI Innovation Strategy (NAIS) and Digital Transformation Centers, though specific details were unclear.

## Infrastructure as Foundation

A central theme throughout the discussion was that meaningful AI implementation cannot occur without addressing basic connectivity challenges. This perspective fundamentally shaped the conversation, with speakers consistently returning to infrastructure requirements.

Lacina Kone’s analysis was particularly clear on this point, noting that the usage gap represents a more significant challenge than the coverage gap. Where internet infrastructure exists, barriers include costs exceeding UN recommendations of 2% of monthly revenue, lack of devices, insufficient local content, and limited digital skills.

Each country representative shared their infrastructure initiatives, though details varied in clarity due to transcript quality. Zimbabwe reported progress in internet penetration with plans for universal coverage. The Philippines mentioned comprehensive broadband programs across their island nation. Lithuania focused on infrastructure resilience and fiber connectivity goals.

## National AI Strategies and Approaches

The discussion revealed diverse approaches to AI strategy development, with each country crafting frameworks reflecting their specific contexts and challenges.

**Zimbabwe** appears to focus heavily on agricultural applications, with mentions of precision farming, weather monitoring, and equipping agricultural extension officers with technology. The country emphasizes bridging urban-rural divides through targeted programs.

**Costa Rica** described collaborative strategy development with over 50 institutions. The country appears to be transitioning from receiving international cooperation to providing knowledge and tools to regional neighbors.

**The Philippines** developed their strategy with strong presidential leadership and multi-agency collaboration. They’ve established specialized centers targeting marginalized sectors and developed skills frameworks in partnership with industry.

**Lithuania** emphasized leveraging small country advantages through flexibility and reduced bureaucracy, proposing regulatory sandbox approaches based on their FinTech experience.

**Smart Africa** launched a continental AI Council addressing computing power, datasets, algorithms, and governance across Africa’s diverse linguistic landscape.

## Skills Development and Digital Inclusion

Skills development emerged as a critical component of bridging the AI divide, with speakers recognizing the need for comprehensive programs addressing diverse population segments.

Lithuania’s categorization of digital natives, immigrants, and refugees provided a useful framework for understanding different population needs. The concept of digital refugees—those intimidated by digital tools—represents both a challenge and an opportunity to discover new participants in the digital economy.

Other countries described various skills initiatives, though specific details were often unclear in the transcript. Zimbabwe mentioned urban-rural bridging programs, Costa Rica highlighted public sector training with high demand, and the Philippines described industry-crafted skills frameworks.

## Sector-Specific Applications

Agriculture emerged as a particularly significant opportunity for AI implementation across developing nations. Zimbabwe’s focus on agricultural AI reflects the sector’s economic importance, while Costa Rica mentioned agricultural laboratories combining multiple technologies.

Several countries are also advancing e-government services through AI integration, though specific applications were not clearly detailed in the available transcript.

## International Cooperation and Knowledge Sharing

The discussion revealed evolving approaches to international cooperation, moving beyond traditional donor-recipient models toward knowledge-sharing partnerships.

Costa Rica’s evolution toward providing tools and assistance to regional neighbors exemplifies this shift. However, Lacina Kone emphasized the importance of technological sovereignty, noting that Africa must develop AI solutions in local languages to avoid “riding in someone else’s future.”

The linguistic challenge is particularly acute in Africa, with over 2,000 languages requiring representation in AI systems—representing perhaps the most complex multilingual AI development challenge globally.

## Regulatory Approaches and Governance

The panel revealed different approaches to AI regulation and governance, though many details were unclear due to transcript quality.

Lithuania advocated for regulatory sandbox approaches with differentiated regulation based on risk levels, leveraging small countries’ flexibility as competitive advantages. Smart Africa described more structured governance through their AI Council, addressing multiple aspects of AI development across diverse national contexts.

## Key Challenges and Opportunities

Several significant challenges emerged from the discussion:

– **Connectivity**: The fundamental challenge of connecting 2.6 billion people globally

– **Affordability**: Internet costs exceeding recommended percentages of income

– **Skills**: Need for comprehensive digital literacy programs

– **Linguistic Diversity**: Developing AI systems that serve local languages and cultures

– **Governance**: Balancing oversight with innovation flexibility

## Forward-Looking Perspectives

*Note: The transcript appears to be truncated, and the concluding portions of the discussion were unclear or incomplete.*

The speakers emphasized the urgency of addressing these challenges, with references to 2030 targets for connectivity and AI development goals. However, specific forward-looking commitments and timelines were not clearly articulated in the available transcript.

## Conclusion

This discussion highlighted that bridging the AI divide requires addressing fundamental digital infrastructure gaps before advancing to sophisticated AI applications. The speakers demonstrated consensus on the importance of inclusive approaches that ensure “no one is left behind,” while recognizing that different countries must develop tailored strategies based on their specific contexts and development stages.

The most significant insight was the recognition that AI development cannot be separated from broader digital development challenges. The emphasis on infrastructure-first approaches, combined with attention to skills development and cultural relevance, provides a framework for equitable AI development globally.

However, significant challenges remain, including connecting billions of unconnected people, ensuring affordability, developing multilingual AI systems, and creating governance frameworks that balance innovation with appropriate oversight. The path forward requires continued collaboration and commitment to inclusive development approaches.

*This summary reflects the content available in the provided transcript, which contained significant portions of unclear or fragmented speech. Some policy details and technical specifications mentioned by speakers could not be clearly verified due to audio quality issues.*

LJ Rich: H.E. Dr. Tatenda Anastasia Mavatera, Minister, Ministry of Information, Communication, Technology, Postal and Courier Services, Zimbabwe H.E. Mr. Hubert Vargas Picardo, Vice Minister, Ministerio de Ciencia, Innovación, Tecnología y Telecomunicaciones, Costa Rica H.E. Jurate Sobhiene, Chair of the Council, the Communications Regulatory Authority of the Republic of Lithuania H.E. Emy-Lou Vassouza Delfin, Director, Department of Information and Communications Technology at the Philippines H.E. Lassina Kone, Director General and Chief Executive Officer, Smart Africa H.E. and is moderated brilliantly by Ittasam Al-Masraei, Executive Director of the Office of AI and Advanced Technology at the Department of Finance H.E. CEO and Founder of AIE3, Chairperson of the UN AI for Good Impact Initiative Ladies and gentlemen and distinguished guests, I’m absolutely certain that you’ll really enjoy this next panel. Thank you.

Ebtesam Almazrouei: Good afternoon, everyone. It’s our pleasure to have you here today with us again and discussing a very important topic. How we can bridge the AI divide across different nations and regional countries. I am delighted and honored today to have Her Excellency, Dr. Tedinda, the Minister of Zimbabwe, Ministry of Information and Communication Technology. His Excellency, Mr. Hubert, Vice Minister, Ministry of Science Innovation Technology in Costa Rica. Jurata, Chair of the Council, the Communications Regulatory Authority of the Republic of Lithuania. Welcome. Emy, the Director of the Department of Information and Communications Technology from Philippines. And Mr. Lesina, Director General and CEO of South Africa. Smart Africa. In today’s session, it is very important to discuss how we can bridge the AI divide as a nation. What are the best practices and approaches we should have and listen from different countries, how they are doing their capacity building model. How they are investing in the infrastructure. What are the best mechanism and the framework to have international cooperations and best implementation so we can bridge the AI divide. I would like to start the session by asking each one of you today. What does bridging the AI divide means for your country or for your regional context? So we start from the left. Thank you very much for inviting me.

Lacina Kone: Before talking about the bridging of AI, bridging the gap of the AI, there are gaps already, digital gap. You have to remember one thing, the AI sits on the broadband. So what do we do first of all to bridge the broadband? Today we have 2.6 billion people who are not connected. So if you want to bridge the AI, you have to start by that first of all. But AI is actually on the outset is much, much bigger. Because traditionally we all knew that we have a digital inclusion problem and digital gap problem. But the AI goes all the way to the cultural. Because now with the rise of the LLM, the rise of the medium-large language and the small-large language, it basically means that even people who have not been to school, they can also access technology. So there’s an impact. The gap becomes bigger. It means it’s more awareness even before. I think it’s a wake-up call actually. And to be able to bridge that, first of all of course you know that the AI sits on computing power. You might say in Africa we don’t have the energy, we don’t have this. How are we going to compete? Okay, we have a data set. The data set, we have 1.4 billion people living in Africa. Young people which represent 70% is actually less than 30 years old. We have algorithm because when you talk about AI, it’s all about bias as well. So how do we overcome this? Because these are all bridges. There is no AI for us success in AI if there’s no collaborations. Because we have over 2,000 languages. It’s our responsibility in Africa to actually train large language models in our language. Otherwise, we’ll be basically riding in someone else’s future which is not ours. And the best way for us, that’s why at Smart Africa we launched the AI Council in April 4th in Kigali. Africa AI Council which comprises of 15 members of the council. Eight coming from the private sector and seven coming from the government sector. Because it has to be private sector driven to look into the computing power, to look into the data set and to look into the algorithm and to look also into the AI governance itself. Because when you talk about AI in the West, you’re talking maybe French, English, Spanish. But we have over 2,000 languages. How do we deal with that? This is how we think that we can bridge the AI gap.

Ebtesam Almazrouei: Thank you. Thank you. And to Her Excellency, what are your views about the AI divide from your perspective and from your country perspective? Thank you very much.

Tatenda Annastacia Mavetera: I really want to thank you. Thank you IITU for giving us this opportunity. And Zimbabwe to come and say our story. When we’re talking about bridging the AI divide, we are making sure that we want to make sure that we’re leaving no one and no place behind. We have got a national philosophy which has been capacitated by our president, His Excellency Dr. Idim Nangagwa, where he has said that when we are going to provide for technology, when we are going to provide for any initiative, we need to make sure that no one is left behind. So we are saying we need to bridge the digital divide, especially between the urban and the rural areas. And we are looking at various initiatives. We have come up again with an acronym, which we are using in Zimbabwe, which we are calling PSL. And PSL is PSL and then I squared. What are we saying? We are looking firstly on P. We are looking on participation. We are looking at privacy. How are people going to participate? That is bridging the digital divide. How are people going to also observe privacy? That’s on the P. On the S, we are talking about skills. How are we going to upskill? Where are we? What is it? Where are the gaps? Do we need to upskill? Do we need to reskill? So it means we need to look for initiatives around skilling. And then we talk about L, which is leadership. Leadership, the political will, how we are going to relate, how we are also going to come up with policies, sound policies, which are conducive and which would also allow for an AI revolution. It’s also very prudent. And we are also looking at L again, which then becomes legislation. As Zimbabwe, we have come up with an artificial intelligence framework and now we have concluded our artificial intelligence strategy. So it means that when we are looking at all these initiatives, let’s find out the legislation, what regulations we need to put in place. And we are looking at AI, local solutions. It’s not only about looking at what everyone else is doing. Let’s customise it. Let’s make it relevant to Zimbabwe. Let’s make sure that all privacy concerns are addressed. And also we need to look at what the Zimbabweans can be able to appreciate, how any person can be able to relate to whatever they have to do. So for us, when we are bridging the digital divide, bridging the AI divide, it’s making sure that no one is left behind. When we are talking about connectivity, when we are talking about connectivity, and be a different case when it comes to how to achieve bacterial and reproductive health. We should not waste resources. We have very limited capacity.

Hubert Vargas Picado: One of the important things that we are trying to do is to solve this problem, and we are trying to do this in a way that is not only for us, but for all of us. Just for context, Costa Rica has an 85% connectivity coverage in Costa Rica, but in terms of use of AI, the percentage is much lower, right? So for us, bridging the ADOI is fundamentally about equity. For us, artificial intelligence is not an end in itself, but it is an end in itself, and we are trying to make sure that we are not only focusing on infrastructure or talent, but we are closing these gaps in terms of socializing values like inclusion, ethical governance, and technological sovereignty. Our national AI strategy that was launched just ten months ago embodies this vision. It was co-created with 50-plus institutions, and it is in line with a multilingual strategy, and my interest as Non-Governoral Executive was to develop in partnership with UNESCO vincent on AI ethics principles. We are prioritizing inclusive for rural communities, indigenous people, and we have important areas to cover in this, and youth specifically. To programs like community innovation and community development, we are working with the UNESCO to develop inclusive for rural communities, and we have more than 200,000 of these community centers. We bring basic AI literacy and problem-solving tools to underserved areas, which have Internet, but the meaningful connectivity now includes AI. So bridging the IDI means ensuring that no one is left behind. You have seen workshops with different

Ebtesam Almazrouei: underserved areas, inclusive leeing programs, and concerned about these needs we’re also working with countries who have diversity in labor market, and yet they have unacceptable rights, and having small resources in Africa is a challenge actually since, we can only live with one LLC for two months, in Europe which is livable for So I am thinking about two prerequisites,

Jurate Soviene: First of all, digit infrastructure and digital skills. Talking about digital infrastructure, so in lithium inia ble you have connect ed by mobile networks, and what our risk is, current our task is to strengthen our infrastructure to make it more resilient and to have higher quality of services, to have a broadband connection, fibre connection to each household by 2030. And here, the extremely important thing is the holistic approach to the whole infrastructure, national infrastructure, not only digital infrastructure. For example, do we plan fibre network when we construct roads? Do we plan fibre networks, broadband networks, when we expand our electricity network? So these things, these issues, we should, those different infrastructures, we should have a holistic approach to all these different infrastructures. And the second, the second prerequisite is digital skills, which is extremely important, and people of this era, digital era, or AI era, they are sorted into three categories, digital natives, those born with digital tools, digital immigrants, it’s us, we find digital tools and digital issues interesting, and find them helpful, and the third group of people, which is quite large, is digital refugees, those who find digital tools scary and not trustworthy, so, and these are our parents, so, and especially in Europe, they comprise, they are a very huge group of society, because we are an ageing society in Europe, and it’s about one-fifth of our population, so what to do, what to do, what, we are talking about bridging the gap, and we are, we see it as a problem, but maybe we should stop seeing it, closing that gap as a problem, but we should find, should see it as an opportunity, as an opportunity to discover, to discover new e-citizens, to discover, for businesses to discover new, new clients, customers, and this is, this is extremely, to discover new employees in an ageing society, so these, these things are extremely, extremely important, and so, what, what, what, what we can do, we should combine forces, and, and, and yes, and grant opportunity for, for people who, who we, whom we consider as digital refugees to, to, to, to not be, not being excluded from, from, from the society, from the life

Ebtesam Almazrouei: we live, digital life. Thank you, and to Amy, I’d like to hear more and understand from you, from your country, Philippines, what is exactly it means for you, digital

Emmy Lou Versoza Delfin: divide? Thank you, hi, first of all, good afternoon, everyone, I’d like to thank ITU for inviting the Philippines here to be part of this panel, it’s really a great honor to be here, yeah, in the Philippine context, when you talk of bridging the AI divide, just like with the other panelists here, it means addressing the growing gaps in access, skills, infrastructure, and opportunities, of course, related to artificial intelligence, so that us in the Philippines, Filipinos, regardless of location, income, or background, can benefit from and participate in the AI-driven digital economy, and you know, the Philippines is composed of more than 7,000 islands, and it’s really about bringing the opportunities, maybe digital literacy, of course, connectivity, to those people outside of the metros, so right now, actually, we have crafted also our national AI innovation of the Philippines, we call it NAIS, and it’s good that, you know, it’s our president that’s actually taking the lead very actively, it’s composed of different government agencies, and of course, together with the and Dr. H.E. Mr. Solly Malatsi. We are trying to work with the private sector to make sure that, you know, addressing the digital divide, particularly the AI divide, should be addressed.

Ebtesam Almazrouei: Thank you. As we can see here, there are many efforts around the words from different countries, from Africa to Philippine, and all of these efforts, we are trying to put the best international cooperation mechanisms and also Elders around the world to missions and pressure and effort. Every community seems to have the strength to help us in deliberation. Enemy even from Fiji, France. After finding the way, we are trying to enable international cooperation and see if we have resources from different societies to actually raise the bills thanks to the collaboration that we are doing. Thank you.

Lacina Kone: When you say African country already, you completely align with us when you talk about when we said our vision is to transform Africa to a single digital market. There’s one thing, it’s to deploy the AI in one nation, in one country, and there’s a different thing when you say you want to integrate it within Africa. So the number one is the collaboration. The collaboration and access to the high-power service, that’s the third thing and again for working in the African geographical area is one special ум.. collaborate in clusters, collaborate in D, the algorithm we’ll talk about, and also all of this should be under the AI governance, a proper AI governance. Of course, it has to be aligned with what the high-level UN talked about, and also by looking at the African Union AI strategy, so if one thing I need to talk about, we need a collaboration across all of these, even if you look at the national level. It is not one size should fit all, but all the size should fit together, you have to look at it, it’s not like only one way to reach to Roma, so when I’m saying this, that also requires a coordination, technology is not the limit, technology is the limit, this is why we are in Beja Keko and Taliban and it was, when I delivered the well I was actually back in 2010 when I delivered the

Ebtesam Almazrouei: meeting when we were in Beja Keko when I started talking, that is when 40% of the nations were like you should and when it becomes too big, it will end up inside the land, but that is the your drains. Yes, small and over regulating is the poisonous mixture I

Jurate Soviene: would say and how to turn our size into advantage. I think that we can share one good example we had in FinTech sector in Lithuania. We became a FinTech hub not because we had big financial institutions but because we made it easy for startups to test their products, their ideas in a safe and supervised environment in so-called regulatory sandboxes and these regulatory sandboxes, they allowed companies to work together with authorities to give, to get real-time feedback and to launch faster into the market and it built trust, so why not to use that example in AI regulation. Yes, in EU we have EU AI Act with main principles and but we know that the majority of AI solutions, they are low-risk solutions, low-risk products and only up to 50% they are high-risk, so we think that it would be a good solution to let self-regulation for those who are low-risk and to observe within the regulatory sandbox those who are high-risk solutions, so we think that this might be the approach and as small country and other small countries we should find our competitive advantage, it is not wise to compete with powerful state superpowers, we should be flexible, open for experimentation, we should be fast and with low bureaucratic bureaucracy and that might be our chance, our opportunity to attract and to become a platform for AI products, for experimentation, for innovations. Okay, excellent and we take it from innovation

Hubert Vargas Picado: and we go to His Excellency, your title has innovation, can you tell us more about what are the best practices to enable the AI innovation in your country? Well, innovation carries a lot of weight but it is by design that we include that word in the ministry and our AI strategy include six initiatives that I’m gonna be short describing it but it is that they are intended to actually close some innovation gaps that we have detected in our country. For example, we’re developing a public sector training program because we detected that there’s an important gap between the private sector and the public sector in terms of accession to AI on its tools and it’s currently a success, actually this week started the first cohort and more than ten people, more than ten times the capacity of the cohort was the interest showed by the public sector and we are intended to open cohorts like on a regular basis. The second gap that we detected is basically the SMEs sector, if you contrast them with the multinationals that are operating in Costa Rica and that’s why we have opened laboratories, for example a 5G testbed with AI applications, we have opened a laboratory that is called AgroBoost to implement artificial intelligence in decision-making in coffee, banana, the most advanced is the coffee program in the Los Santos area in Costa Rica that is combining 5G, IoT and AI and it is marvelous and we intend to escape from that pilot and in terms of program, in terms of how even to, on how to innovate with neighbors, we developed the OECD toolkit, AI toolkit just like one month ago and our intention with that toolkit was not only to position ourselves in the international spectrum but also to help neighbors, Central America, the Caribbean, we are now a high-income country and we have to change our mindset, we are not receptors of cooperation anymore, we have to be more friendly with our knowledge and that’s That’s why that’s in some how we project to be more innovative, not only particularly in AI implementation, but also changing our relationship with our neighbors. Okay, great.

Ebtesam Almazrouei: And to Amy, we know that the Philippines is made up of thousands of islands, and maybe it is very hard to put an AI strategy. What is your approach in that, how you can implement the proper AI strategy across different islands? Of course, we have to be more inclusive, you know, we have unique challenges. Because I mentioned we have more than 7,000 islands in the country.

Emmy Lou Versoza Delfin: So as I mentioned also that we have developed our national AI innovation strategy. And of course, to be able to have a future ready AI, we should have a robust connectivity. So having said that, the Philippines, actually, as I speak here, our Philippine president actually launched our phases two and three in terms of the National Broadband Program. We’re done with our phase one. But basically, what we’re trying to say is we have to make sure that we have really strong connectivity. And aside from the National Broadband Program, we’re also implementing the free Wi-Fi for all, because we have our GIDAs, or geographically isolated and disadvantaged areas in the country. So we’re implementing the free Wi-Fi for all program. So that means giving free connectivity to public places, public schools, health centers, public plazas, elementary and high school. And yeah, so those public places in those really last class municipality in the Philippines. So we’re doing that. Right now, we have 17,000 free Wi-Fi sites in the country. But of course, we need to do more. Currently, you know, we have 117 million population in the Philippines, 73% are internet users. So we have to make sure that, yeah, all of them, regardless of their location, are digitally connected. So that’s connectivity. In terms of digital upskilling, we have developed the Philippine Skills Framework for AI analytics. It’s actually crafted by the industry to make sure that our graduates are ready to take on AI and analytics-related jobs. But of course, we have to make sure that, because of our challenges in terms of, yeah, of our archipelagic location, so we have to be more inclusive. And there’s also another initiative that we have in the Philippines, actually patterned after ITU. So the Digital Transformation Centers, or Technology for Education Centers, so that our marginalized sectors, out-of-school youth, persons with disabilities, senior citizens, can actually just go to the Digital Transformation Centers to make sure that, you know, we have the facility for free in terms of equipment and connectivity. But of course, there are a lot of more things that we have to do to make sure that we have to cover all the 7,100 islands that I’ve mentioned.

Ebtesam Almazrouei: I’m sure it is a tough work to do, and you are up to that challenge, Amy. Your Excellency, Dr. Tatenda, what is Zimbabwe’s biggest opportunity today in the AI divide, and how you are trying to bridge that gap, and what is the opportunity there that you see today?

Tatenda Annastacia Mavetera: Thank you very much. I really want to appreciate that. I’m spoiled for choice. We have got quite two big opportunities that we have in Zimbabwe. The first one is agriculture. Agriculture, Zimbabwe is a country that has got agriculture being its greatest pillar. Even through our development blueprint, which we call NDS1, agriculture is the backbone of our economy. So there is a lot that we really want to appreciate that has been done together with ITU. Firstly, we have got a partnership that we have with the UNDP, which has seen us install 30 automatic weather stations across 13 districts of our country. And also, we have managed to also install 10 automatic rain gauges, and many more other products that have come. Zimbabwe also has been selected to be the lead country under the Smart Africa Initiative after we were selected during our Transform Africa in Zimbabwe, which was hosted in Zimbabwe. So Zimbabwe has been doing a lot of initiatives around AI, using AI for agriculture. We also talk about precision farming. We have been able to look at the use of drones, artificial intelligence, for crop monitoring, for soil analysis, and also pest control. As Zimbabwe, we have got 6,000 agri-tech extension officers, and we have capacitated them. All of them, we have given them electronic gadgets. We have also powered them with the use of low-Earth-orbit satellites, having a styling kit. We have been able to do that. We have realized that in our rural areas, this is where our agri-tech extension officers are more. So there isn’t much connectivity there. We have been having a challenge because of that. So we realized that there is a need for us to adopt the low-Earth-orbit satellites, which make them at least be able to provide for real-time advice to farmers via their various platforms, including the mobile platforms. As Zimbabwe, again, we have realized that now we have started having prototypes of creating our own drones in our universities, because there is a great opportunity for agriculture in Zimbabwe. Of course, secondly, I’ll talk about e-government services. We have partnered with quite a lot of countries in us having an investment in enterprise architecture modeling. And now, as Zimbabwe, we are now advancing in terms of our e-services. This is also another great opportunity. We need to provide for good services to the people of Zimbabwe, and this can be done by us using artificial intelligence. And as Zimbabwe, again, we have come up with what we call digital centers. This is how we have been able to integrate and make sure that we will make each and every other person be able to access digital services. So these digital centers, they are in each and every district in Zimbabwe, and this is where our farmers can be able to access market processes. Whenever they want to market their products, they can be able to access them at our digital centers. And finally, as Zimbabwe as well, we have also, the same way which has been said by the Philippines, we also have got a free Internet program whereby we have got free public Wi-Fi spaces to make sure that everyone has got access to the Internet. We believe that when we’re talking about artificial intelligence, we’re also talking about the capacity that every person has, and we’re talking about how we can also access electronic gadgets which will be able to, for us, be able to utilize artificial intelligence. So in Zimbabwe, as much as we are very happy through various partnerships, we have really migrated. I’ll tell you, in 2005, we were talking about 20 plus today here at WSIS, but you realize that in 2005, we’re talking about an Internet penetration rate of 5%. But now, 20 years later, we’re talking of an Internet penetration rate which is close to 80%. And what we want to do as Zimbabwe is for us to have 100% Internet penetration rate, and also increase on our mobile penetration rate, and make sure everyone has got access to devices. So it’s Internet for all, and we’re talking about devices for all. These are the two initiatives that we’re doing in line with our presidential Internet scheme, where our president, his excellency, Dr. Edi Mnangagwa, has really assisted us to make sure that everyone has got access to the Internet. So these are initiatives that I think are quite, is an opportunity for us using artificial intelligence, and let me conclude again and say, we had a conversation in the morning where we had our luncheon, where we were saying, AI is not going to replace jobs, but tasks are going to be replaced by AI. So it’s not about the job, but it’s about you making sure that the task that you’re having is the one which is going to integrate AI, and then it becomes different. I so submit. Thank you. Thank you, Your Excellency. And definitely during the AI governance day in the lunch, it was an interesting discussion

Ebtesam Almazrouei: to hear from different industry leaders and governments and NGOs. , and we’re going to talk about how we can as different nations entities collaborate so we can bridge the A.I. divide. Still, until the date, there is still 2.6 billion people around the world remains offline. Still there are some countries and rural communities even, they lack access to electricity. So, we’re going to talk about how we can bridge the A.I. divide. So, I’m going to start with you, Mr. Bosun, and then I’m going to turn to Mr.

Lacina Kone: H.E. Mr. Solly Malatsi to talk about the A.I. divide. So, I’m going to start with you, Mr. Bosun, and then I’m going to turn to Mr. H.E. Mr. Solly Malatsi to talk about the A.I. But I’m also going to point to the technology, of the technology that we’re using in Africa, when you’re living in African countries, perhaps also one of the reasons that you are so responsible, 2.6 billion that we need to cover from now, five years. That is a big challenge. One thing we all omit, when you look at the continent in Africa, it is not because of the lack of coverage. By the way, the coverage today gap is around 15 to 20 per cent in terms of covering the population. We have something called usage gap, which is killing our continent. It means that the infrastructure exists in the area, the broadband, mobile broadband is deployed, but people are not using it for four reasons. Affordability, the cost of the internet, by the UN recommendations, the internet should not be more expensive than 2 per cent of your monthly revenue. Two, smart devices. Average cost of smart devices in Africa, should I choose to have a phone or to pay my rent? Third thing, local content, why am I using the broadband? So this has to be addressed by our start-up. Number four is cyber hygiene. All of these actually are policy levers for AI. The infrastructure, there is no AI if there is no meaningful connectivity for our population. Because, guess what, we talk about digital divide today in 2025. 2020, so we’ve been talking about it for the past 15 years. Now, we’re talking about AI divide. You might well be educated, you might well be connected, if you don’t know how to use the AI, you’re still considered to be illiterate. So, more we push the urgency of connecting everyone based on the UN Broadband Commission recommendations, more we push those decisions, more challenges will be coming from now to 2030. So, this is the call for all of us. In this room here, time is not on our side. If you look at what we achieved in the past 15 years, in the remaining five years to cover, one-third of the planet’s population. Thank you. So, looking one year ahead from now, what would success in bridging the AI divide look like to you? One word or one sentence? One year from now.

Ebtesam Almazrouei: One year from now. One year from now. One year from now, what would the success look like to you in bridging the AI divide?

Lacina Kone: I’m a true believer in a decentralized AI. I do believe in the decentralized. You will be seeing a lot of proliferation around the world. I do believe that we are the first country in the world to have a large language model, the reason is very simple. China, for example, did not have to wait to speak English to be where they are today. What I’m seeing today, you will be very surprised, particularly what will be coming from Africa in terms of localized large language model in the AI. That’s what I believe in.

Ebtesam Almazrouei: Thank you. And your excellency, Dr. Ndung’u, what do you think about the future of AI in Africa? What do you think about the future of AI in Africa?

Tatenda Annastacia Mavetera: Thank you very much. One year from now, I think a lot would have happened. Everyone would be able to understand everything around AI. We’re talking about infrastructure, what is needed. Definitely, we have a lot of people knowing exactly what needs to be done. But, I think, for me, I think we’re going to be able to do a lot more. I think that the skills that I had are not relevant now, so definitely you really need to be up-scaling, re-scaling, or even scaling yourself. So all this is definitely going to be done around that. Again, I would say one year from now, for me, I see a lot of opportunity, innovation. I see a complete shift from what we’re doing now, the technology, to what we’re doing now. I think that the future of AI will be much more about how we can enhance, become efficient,

Ebtesam Almazrouei: reliable. Everything will be so tech-related, so I’m happy the tech space would have gone quite, would have done quite a lot, and I’m sure that is going to be the same. Thank you. Thank you. Your Excellency?

Hubert Vargas Picado: I think that the future is going to be a lot more about how we can enhance, how we can become more efficient, and I respect that we will have as a world a better AI, multilingual one, for example, more multilingual, reduced biases, for example, in gender, the data that we have on that, that’s an important challenge to reduce. And also more integration, because this is a big challenge for all of us, but we need to be more inclusive, more inclusive, more inclusive, and more decision, joint decision-making. So, that’s my approach. Thank you.

Jurate Soviene: A senior in a small town knows very well the place where he can find, he or she can find assistance on using digital tools, because he or she knows that there’s not so much access, not so much tools in Europe, and it doesn’t hurt everybody to take them. But in Europe it’s not possible to prove, but I consider it every is the solution to become a person with access to the grid, so people should have access to the grid, and it’s not the solution to become a person with access to the grid. So, I think we should, by ourselves, we should put more

Ebtesam Almazrouei: effort in writing our data in our own languages, because the language that we write it today, it would be the language of AI. So, it is not the engineers’ fault, it is our fault, and we have to take responsibility and accountability for that. To Amy. Thank you very much.

Emmy Lou Versoza Delfin: So, we have a lot of work to do. We have a lot of work to do on the implementation of the strategy, will be, of course, started and will be continued, more much-improved connectivity in the red areas, more Filipinos, millions of Filipinos will be upskilled, of course, using our skills framework done by the industry, and more use cases for our region filed can increase the construir

Ebtesam Almazrouei: projections for future AI. There is no room to do a processus in the region for AI to be as well, and we are trying, because we need to migrate to future states, and along with the GDPR, we need to meg OOS, send the report into governments, because in the region, the HSU is quantities of capacity for new modern climate models of all nations. That’s it, thanks, Takashi, thank you ever so much to our panellists, helping us understand where the levers might be to move forward. Climate and environmental change ourselves has become the Western culture, I started the conference the other day, and you started where, the climate community?

Agreement points

Infrastructure as Foundation for AI

Speakers

– Lacina Kone
– Tatenda Annastacia Mavetera
– Emmy Lou Versoza Delfin
– Jurate Soviene

Arguments

AI sits on broadband infrastructure, with 2.6 billion people still unconnected globally

Zimbabwe increased internet penetration from 5% in 2005 to nearly 80% today, aiming for 100% coverage

Philippines has 73% internet users among 117 million population, implementing National Broadband Program and free Wi-Fi for all in public places

Lithuania has near-universal mobile network coverage but focuses on strengthening infrastructure resilience and fiber connections to households by 2030

Summary

All speakers agree that robust digital infrastructure is the fundamental prerequisite for AI implementation, with each country working toward universal connectivity through various broadband and Wi-Fi initiatives

Topics

Infrastructure | Development

No One Left Behind Philosophy

Speakers

– Tatenda Annastacia Mavetera
– Hubert Vargas Picado
– Emmy Lou Versoza Delfin
– Jurate Soviene

Arguments

Zimbabwe’s ‘no one left behind’ philosophy requires bridging urban-rural digital divide through targeted skilling initiatives

Costa Rica launched national AI strategy co-created with 50+ institutions, emphasizing equity, inclusion, and ethical governance

Philippines developed AI analytics skills framework crafted by industry and established Digital Transformation Centers for marginalized sectors

Lithuania categorizes population into digital natives, immigrants, and refugees, viewing digital inclusion as opportunity rather than problem

Summary

All speakers emphasize inclusive approaches to AI development that specifically target marginalized communities, rural populations, and vulnerable groups to ensure equitable access to AI benefits

Topics

Development | Human rights

Skills Development and Capacity Building

Speakers

– Lacina Kone
– Tatenda Annastacia Mavetera
– Hubert Vargas Picado
– Emmy Lou Versoza Delfin

Arguments

Need to address skills gaps through upskilling and reskilling programs, with focus on making AI accessible to those without formal education

Zimbabwe developed PSL framework focusing on Participation, Privacy, Skills, Leadership, and Legislation with completed AI strategy

Costa Rica identified gaps between private and public sectors, launching training programs with high demand from public sector workers

Philippines developed AI analytics skills framework crafted by industry and established Digital Transformation Centers for marginalized sectors

Summary

All speakers recognize the critical need for comprehensive skills development programs that address both technical AI literacy and broader digital skills across different sectors and populations

Topics

Development | Sociocultural

National AI Strategies and Frameworks

Speakers

– Tatenda Annastacia Mavetera
– Hubert Vargas Picado
– Emmy Lou Versoza Delfin

Arguments

Zimbabwe developed PSL framework focusing on Participation, Privacy, Skills, Leadership, and Legislation with completed AI strategy

Costa Rica launched national AI strategy co-created with 50+ institutions, emphasizing equity, inclusion, and ethical governance

Philippines crafted National AI Innovation Strategy (NAIS) with presidential leadership and multi-agency collaboration

Summary

All three countries have developed comprehensive national AI strategies with strong government leadership, multi-stakeholder involvement, and focus on ethical, inclusive implementation

Topics

Legal and regulatory | Development

Similar viewpoints

Both speakers emphasize the critical importance of developing AI systems in local languages to maintain cultural relevance and technological sovereignty, with responsibility placed on communities to actively create content in their native languages

Speakers

– Lacina Kone
– Ebtesam Almazrouei

Arguments

Africa must train large language models in local languages to avoid ‘riding in someone else’s future’

Importance of writing data in local languages since ‘the language we write today will be the language of AI’

Topics

Sociocultural | Legal and regulatory

Both countries advocate for flexible, experimental approaches to AI regulation and innovation that allow for testing and development in controlled environments while maintaining appropriate oversight

Speakers

– Jurate Soviene
– Hubert Vargas Picado

Arguments

Lithuania advocates for regulatory sandboxes for AI innovation, allowing self-regulation for low-risk solutions while supervising high-risk ones

Costa Rica established 5G testbed and laboratories to bridge innovation gaps between SMEs and multinationals

Topics

Legal and regulatory | Economic

Both countries prioritize agricultural applications of AI as key economic opportunities, implementing comprehensive technology solutions that combine AI with other advanced technologies for agricultural decision-making and productivity improvement

Speakers

– Tatenda Annastacia Mavetera
– Hubert Vargas Picado

Arguments

Zimbabwe focuses on agriculture as backbone of economy, using AI for precision farming, crop monitoring, and equipping 6,000 agri-tech extension officers with technology

Costa Rica developed AgroBoost laboratory combining 5G, IoT, and AI for coffee and banana production decision-making

Topics

Development | Infrastructure

Unexpected consensus

Small Countries as Innovation Leaders

Speakers

– Jurate Soviene
– Hubert Vargas Picado

Arguments

Small countries should leverage flexibility and low bureaucracy as competitive advantages rather than competing with superpowers

Costa Rica shifted from cooperation recipient to knowledge provider, developing OECD AI toolkit to help Central American and Caribbean neighbors

Explanation

Unexpectedly, both Lithuania and Costa Rica demonstrate that smaller nations can become AI innovation leaders by leveraging their inherent advantages of agility and reduced bureaucracy, rather than viewing their size as a limitation. This challenges conventional assumptions about AI development requiring large-scale resources

Topics

Legal and regulatory | Economic

Usage Gap More Critical Than Coverage Gap

Speakers

– Lacina Kone

Arguments

Coverage gap is only 15-20% in Africa, but usage gap is the real problem due to affordability, device costs, local content, and cyber hygiene issues

Explanation

Kone’s insight that Africa’s main challenge is not lack of infrastructure but rather barriers to using existing infrastructure represents an unexpected consensus point that reframes the digital divide discussion from infrastructure deployment to accessibility and usability factors

Topics

Infrastructure | Development | Economic

Overall assessment

Summary

The speakers demonstrate remarkable consensus on fundamental principles of AI development including infrastructure-first approaches, inclusive implementation, comprehensive skills development, and the need for national strategies with strong government leadership

Consensus level

High level of consensus with strong alignment on core principles and approaches. The implications suggest that despite different geographical and economic contexts, there are universal best practices for bridging the AI divide that can be adapted across different countries and regions. This consensus provides a solid foundation for international cooperation and knowledge sharing in AI development.

Different viewpoints

Priority approach to bridging AI divide – infrastructure first vs. holistic approach

Speakers

– Lacina Kone
– Jurate Soviene

Arguments

AI sits on broadband infrastructure, with 2.6 billion people still unconnected globally

Lithuania has near-universal mobile network coverage but focuses on strengthening infrastructure resilience and fiber connections to households by 2030

Summary

Kone emphasizes that basic connectivity must be established first before addressing AI, arguing you cannot have AI without broadband. Soviene, from a country with near-universal coverage, focuses on infrastructure quality and resilience rather than basic connectivity, suggesting different priorities based on development stage.

Topics

Infrastructure | Development

Regulatory approach – centralized governance vs. decentralized flexibility

Speakers

– Lacina Kone
– Jurate Soviene

Arguments

Smart Africa launched AI Council with 15 members to address computing power, datasets, algorithms, and AI governance across 2,000+ African languages

Lithuania advocates for regulatory sandboxes for AI innovation, allowing self-regulation for low-risk solutions while supervising high-risk ones

Summary

Kone advocates for coordinated governance through formal AI councils and structured approaches, while Soviene promotes flexible regulatory sandboxes with self-regulation for low-risk AI solutions. This reflects different philosophies about the role of formal governance structures.

Topics

Legal and regulatory | Economic

Scale of AI development – continental integration vs. national focus

Speakers

– Lacina Kone
– Tatenda Annastacia Mavetera
– Emmy Lou Versoza Delfin

Arguments

Need for decentralized AI development with localized large language models, particularly from Africa

Zimbabwe developed PSL framework focusing on Participation, Privacy, Skills, Leadership, and Legislation with completed AI strategy

Philippines crafted National AI Innovation Strategy (NAIS) with presidential leadership and multi-agency collaboration

Summary

Kone emphasizes continental-scale AI development and integration across Africa, while Mavetera and Delfin focus primarily on national-level strategies and frameworks. This represents different approaches to the scale and scope of AI development coordination.

Topics

Legal and regulatory | Development

Unexpected differences

Role of small countries in AI development

Speakers

– Jurate Soviene
– Hubert Vargas Picado

Arguments

Small countries should leverage flexibility and low bureaucracy as competitive advantages rather than competing with superpowers

Costa Rica shifted from cooperation recipient to knowledge provider, developing OECD AI toolkit to help Central American and Caribbean neighbors

Explanation

Both speakers represent smaller countries but have fundamentally different approaches to their role in AI development. Soviene advocates for small countries to avoid competing with superpowers and instead leverage their flexibility, while Picado describes Costa Rica’s transformation into a knowledge provider and regional leader. This unexpected disagreement reveals different philosophies about whether small countries should focus on niche advantages or aspire to leadership roles.

Topics

Legal and regulatory | Economic

Overall assessment

Summary

The panel showed moderate disagreement primarily around implementation approaches rather than fundamental goals. Key areas of disagreement included infrastructure priorities (basic connectivity vs. quality enhancement), regulatory approaches (structured governance vs. flexible sandboxes), development scale (continental vs. national focus), and the role of smaller countries in AI development.

Disagreement level

The disagreement level is moderate and constructive, reflecting different development stages and national contexts rather than fundamental philosophical differences. These disagreements are actually beneficial as they demonstrate diverse pathways to achieving similar goals of bridging the AI divide. The implications suggest that successful AI development requires tailored approaches based on each country’s specific circumstances, infrastructure maturity, and regional context, rather than a one-size-fits-all solution.

Partial agreements

Similar viewpoints

Both speakers emphasize the critical importance of developing AI systems in local languages to maintain cultural relevance and technological sovereignty, with responsibility placed on communities to actively create content in their native languages

Speakers

– Lacina Kone
– Ebtesam Almazrouei

Arguments

Africa must train large language models in local languages to avoid ‘riding in someone else’s future’

Importance of writing data in local languages since ‘the language we write today will be the language of AI’

Topics

Sociocultural | Legal and regulatory

Both countries advocate for flexible, experimental approaches to AI regulation and innovation that allow for testing and development in controlled environments while maintaining appropriate oversight

Speakers

– Jurate Soviene
– Hubert Vargas Picado

Arguments

Lithuania advocates for regulatory sandboxes for AI innovation, allowing self-regulation for low-risk solutions while supervising high-risk ones

Costa Rica established 5G testbed and laboratories to bridge innovation gaps between SMEs and multinationals

Topics

Legal and regulatory | Economic

Both countries prioritize agricultural applications of AI as key economic opportunities, implementing comprehensive technology solutions that combine AI with other advanced technologies for agricultural decision-making and productivity improvement

Speakers

– Tatenda Annastacia Mavetera
– Hubert Vargas Picado

Arguments

Zimbabwe focuses on agriculture as backbone of economy, using AI for precision farming, crop monitoring, and equipping 6,000 agri-tech extension officers with technology

Costa Rica developed AgroBoost laboratory combining 5G, IoT, and AI for coffee and banana production decision-making

Topics

Development | Infrastructure

Key takeaways

Bridging the AI divide requires first addressing the digital divide, as AI depends on broadband infrastructure with 2.6 billion people still unconnected globally

The usage gap is more critical than coverage gap – infrastructure exists but people don’t use it due to affordability, device costs, lack of local content, and cyber hygiene issues

National AI strategies must be inclusive and culturally relevant, with emphasis on ‘leaving no one behind’ and addressing urban-rural divides

Localization is crucial – countries must develop AI solutions in their own languages rather than relying on external models to avoid ‘riding in someone else’s future’

Skills development through upskilling and reskilling programs is essential, with focus on making AI accessible across all education levels and age groups

Sector-specific applications like agriculture and e-government services offer significant opportunities for AI implementation in developing countries

Small countries should leverage flexibility and low bureaucracy as competitive advantages rather than trying to compete directly with superpowers

International cooperation should shift from traditional donor-recipient models to knowledge-sharing partnerships between countries at different development stages

Regulatory approaches should balance innovation with governance, using models like regulatory sandboxes for safe experimentation

Resolutions and action items

Smart Africa launched AI Council with 15 members to address computing power, datasets, algorithms, and AI governance across African languages

Countries committed to implementing their national AI strategies with specific timelines (Costa Rica’s 10-month-old strategy, Philippines’ NAIS implementation)

Zimbabwe to continue expanding digital centers to every district and achieve 100% internet penetration

Philippines to continue phases two and three of National Broadband Program and expand free Wi-Fi sites beyond current 17,000 locations

Lithuania to achieve fiber connection to each household by 2030 and strengthen infrastructure resilience

Costa Rica to scale up successful pilot programs like AgroBoost laboratory and expand public sector AI training cohorts

Countries to focus on writing and developing content in local languages to ensure AI represents their cultures and needs

Unresolved issues

How to achieve the ambitious goal of connecting 2.6 billion people in the remaining five years to 2030

Addressing the fundamental affordability challenge where internet costs exceed UN recommendations of 2% of monthly revenue

Balancing the need for AI governance and regulation while maintaining innovation and flexibility

Managing the challenge of developing AI solutions across thousands of languages (2,000+ in Africa alone)

Ensuring meaningful connectivity that includes AI capabilities, not just basic internet access

Addressing the digital refugee population, particularly aging societies in Europe who find digital tools scary and untrustworthy

Scaling successful pilot programs and innovations to national and regional levels

Coordinating international cooperation mechanisms effectively across different development stages and regional contexts

Suggested compromises

Adopting a holistic approach to infrastructure development that integrates digital, road, and electricity networks rather than building them separately

Using regulatory sandboxes that allow self-regulation for low-risk AI solutions while maintaining supervision for high-risk applications

Viewing digital inclusion challenges as opportunities to discover new citizens, customers, and employees rather than just problems to solve

Shifting from seeing AI as job replacement to task enhancement, focusing on integrating AI into existing work rather than eliminating positions

Balancing centralized AI governance principles with decentralized, localized AI development to respect cultural and linguistic diversity

Combining public and private sector efforts in AI development rather than relying solely on government or market-driven approaches

Before talking about the bridging of AI, bridging the gap of the AI, there are gaps already, digital gap. You have to remember one thing, the AI sits on the broadband. So what do we do first of all to bridge the broadband? Today we have 2.6 billion people who are not connected. So if you want to bridge the AI, you have to start by that first of all.

Speaker

Lacina Kone

Reason

This comment reframes the entire discussion by establishing a foundational hierarchy of needs – you cannot address AI divides without first addressing basic connectivity. It challenges the assumption that AI solutions can be implemented without addressing fundamental infrastructure gaps.

Impact

This comment set the tone for the entire discussion by grounding it in practical reality. It influenced subsequent speakers to address infrastructure and connectivity as prerequisites, with multiple panelists later referencing connectivity challenges and broadband programs in their responses.

But the AI goes all the way to the cultural. Because now with the rise of the LLM, the rise of the medium-large language and the small-large language, it basically means that even people who have not been to school, they can also access technology. So there’s an impact. The gap becomes bigger.

Speaker

Lacina Kone

Reason

This insight reveals a paradox – AI can democratize access to technology for the uneducated, but simultaneously widens the gap for those without access. It introduces the cultural dimension of AI divides, moving beyond technical considerations to societal impact.

Impact

This comment elevated the discussion from technical infrastructure to cultural and educational implications, leading other speakers to address language diversity, local content, and cultural relevance in their AI strategies.

We have over 2,000 languages. It’s our responsibility in Africa to actually train large language models in our language. Otherwise, we’ll be basically riding in someone else’s future which is not ours.

Speaker

Lacina Kone

Reason

This powerful metaphor about ‘riding in someone else’s future’ crystallizes the sovereignty issue in AI development. It highlights how language representation in AI models is fundamentally about self-determination and cultural preservation.

Impact

This comment introduced the critical theme of technological sovereignty that resonated throughout the discussion. Later speakers referenced multilingual approaches and local solutions, with Hubert mentioning multilingual AI as a future goal.

We think that it would be a good solution to let self-regulation for those who are low-risk and to observe within the regulatory sandbox those who are high-risk solutions… we should be flexible, open for experimentation, we should be fast and with low bureaucratic bureaucracy and that might be our chance, our opportunity to attract and to become a platform for AI products.

Speaker

Jurate Soviene

Reason

This comment offers a nuanced regulatory approach that turns perceived disadvantages (small size) into competitive advantages. It challenges the one-size-fits-all regulatory approach and proposes differentiated governance based on risk levels.

Impact

This shifted the conversation toward practical governance solutions and influenced the discussion about how smaller nations can compete with superpowers through agility rather than resources.

Maybe we should stop seeing it, closing that gap as a problem, but we should find, should see it as an opportunity, as an opportunity to discover, to discover new e-citizens, to discover, for businesses to discover new, new clients, customers… to discover new employees in an ageing society.

Speaker

Jurate Soviene

Reason

This reframes the digital divide from a deficit model to an opportunity model, particularly regarding older populations. It challenges the assumption that digital refugees are simply problems to be solved rather than untapped potential.

Impact

This perspective shift influenced how other panelists discussed inclusion, moving from viewing underserved populations as burdens to seeing them as opportunities for growth and innovation.

One thing we all omit, when you look at the continent in Africa, it is not because of the lack of coverage… We have something called usage gap, which is killing our continent. It means that the infrastructure exists in the area, the broadband, mobile broadband is deployed, but people are not using it for four reasons. Affordability, the cost of the internet… smart devices… local content… cyber hygiene.

Speaker

Lacina Kone

Reason

This comment reveals a critical misconception – that connectivity problems are primarily about infrastructure deployment rather than usage barriers. It provides a sophisticated analysis of why infrastructure alone doesn’t solve digital divides.

Impact

This insight redirected the conversation from infrastructure deployment to usage enablement, influencing how other speakers discussed their connectivity programs and highlighting the complexity of digital inclusion beyond mere access.

AI is not going to replace jobs, but tasks are going to be replaced by AI. So it’s not about the job, but it’s about you making sure that the task that you’re having is the one which is going to integrate AI.

Speaker

Tatenda Annastacia Mavetera

Reason

This nuanced distinction between job replacement and task transformation provides a more sophisticated understanding of AI’s impact on employment, moving beyond simplistic fears to practical adaptation strategies.

Impact

This comment helped frame AI as a tool for augmentation rather than replacement, influencing the overall tone of the discussion toward practical integration rather than existential concerns about AI’s impact.

Overall assessment

These key comments fundamentally shaped the discussion by establishing a hierarchy of needs (connectivity before AI), introducing the sovereignty dimension (cultural and linguistic self-determination), reframing challenges as opportunities, and providing nuanced analyses of complex issues like the usage gap and job transformation. The most impactful contributions came from Lacina Kone, who consistently grounded the discussion in African realities while offering sophisticated frameworks for understanding digital divides. The comments collectively moved the conversation from abstract policy discussions to concrete, culturally-aware strategies that acknowledge both the opportunities and complexities of bridging AI divides. The discussion evolved from technical infrastructure concerns to encompass cultural sovereignty, regulatory innovation, and human-centered approaches to AI implementation.

How do we deal with over 2,000 languages in Africa when training large language models?

Speaker

Lacina Kone

Explanation

This addresses the critical challenge of linguistic diversity in AI development and the need for culturally relevant AI solutions across African languages

How are we going to compete in AI when Africa doesn’t have adequate energy and computing power?

Speaker

Lacina Kone

Explanation

This highlights the fundamental infrastructure challenges that need to be addressed before AI can be effectively deployed across Africa

How do we overcome AI bias, especially when we have such diverse datasets?

Speaker

Lacina Kone

Explanation

This is crucial for ensuring AI systems are fair and representative of diverse populations and use cases

How do we plan fiber network integration when constructing roads and expanding electricity networks?

Speaker

Jurate Soviene

Explanation

This addresses the need for holistic infrastructure planning to maximize efficiency and reduce costs in digital infrastructure deployment

How do we help digital refugees (those who find digital tools scary) become included in digital society?

Speaker

Jurate Soviene

Explanation

This is important for ensuring no population segment is left behind in the digital transformation, especially aging populations in Europe

How can we address the usage gap where broadband infrastructure exists but people aren’t using it due to affordability, lack of smart devices, local content, and cyber hygiene issues?

Speaker

Lacina Kone

Explanation

This identifies key policy levers that need to be addressed to bridge the digital divide before tackling the AI divide

How can we cover one-third of the planet’s population (2.6 billion people) in the remaining five years to 2030?

Speaker

Lacina Kone

Explanation

This highlights the urgency and scale of the connectivity challenge that underlies AI accessibility

How do we ensure meaningful connectivity that includes AI, not just basic internet access?

Speaker

Hubert Vargas Picado

Explanation

This addresses the evolution of connectivity requirements as AI becomes more prevalent in digital services