WS #214 AI Readiness in Africa in a Shifting Geopolitical Landscape

Summary

This discussion focused on AI readiness in Africa within a shifting geopolitical landscape, hosted by the German Federal Ministry for Economic Cooperation and Development with representatives from government, civil society, and private sector across multiple African countries. The session addressed how African nations can build sovereign and resilient AI ecosystems while avoiding digital neocolonialism and ensuring AI serves local needs rather than external interests.

Key speakers emphasized that AI is already impacting governance in Africa through both beneficial applications and harmful uses like automated disinformation campaigns and surveillance tools. Government representatives from Mauritania and South Africa shared their experiences developing national AI strategies, highlighting challenges including limited infrastructure in rural areas, capacity gaps, and coordination difficulties across sectors. Smart Africa’s leadership stressed the need for deliberate, intentional action rather than passive optimism, noting that over 1,000 African startups currently rely on foreign AI models, creating dependency risks.

The discussion revealed that 19 African countries have developed national AI strategies, but implementation remains challenging. Speakers emphasized the importance of inclusive, transparent governance frameworks grounded in human rights and constitutional values. Civil society representatives highlighted their crucial role as watchdogs, educators, and advocates for marginalized communities, stressing the need for mandatory public interest impact assessments and regional coalitions.

Private sector perspectives emphasized creating enabling environments and moving from “paper to pavement” in policy implementation. The conversation underscored Africa’s potential to define its own AI race focused on usefulness rather than raw power, particularly in agriculture, healthcare, and education. Speakers concluded that success requires blended financing, multi-stakeholder participation, and policies that foster homegrown innovation while protecting citizen rights and digital sovereignty.

Keypoints

## Major Discussion Points:

– **AI Governance and Policy Frameworks**: The need for African countries to develop inclusive, transparent AI governance frameworks that are grounded in local values, human rights, and constitutional principles rather than importing wholesale benchmarks from other regions. Multiple speakers emphasized the importance of harmonizing national AI strategies across the continent while respecting individual country sovereignty.

– **Digital Sovereignty and Data Control**: Concerns about AI-driven digital neocolonialism and the risk of African data being processed and trained on models located outside the continent. Speakers stressed the importance of keeping African data in African hands and developing locally-rooted AI solutions that serve African needs rather than external interests.

– **Infrastructure and Capacity Building**: The critical challenges of limited computing power, inadequate internet coverage (especially in rural areas), digital literacy gaps, and the need for significant investment in technical expertise and infrastructure to support AI development across Africa.

– **Multi-stakeholder Collaboration and Inclusivity**: The importance of involving diverse stakeholders – governments, private sector, civil society, youth, and local communities – in AI development and governance, with particular emphasis on democratizing AI literacy and ensuring marginalized voices are heard in policy-making processes.

– **Practical Implementation and Funding**: Moving beyond policy documents to actual implementation, with discussions about blended financing models, public-private partnerships, and the need for realistic, actionable steps that governments can take to build AI readiness while addressing youth concerns about job displacement.

## Overall Purpose:

The discussion aimed to explore how African countries can build sovereign and resilient AI ecosystems that strengthen rather than undermine democratic governance, while navigating a complex geopolitical landscape marked by technological rivalry and economic pressures. The session sought to identify practical strategies for ensuring AI serves local African needs and values rather than perpetuating digital colonialism.

## Overall Tone:

The discussion maintained a constructive and collaborative tone throughout, characterized by cautious optimism balanced with realistic acknowledgment of significant challenges. Speakers demonstrated both urgency about the need for action and pragmatism about Africa’s current limitations. The tone was notably inclusive and pan-African in perspective, with participants building on each other’s points rather than disagreeing. There was a consistent thread of determination that Africa can succeed in AI development by defining its own terms and leveraging its unique strengths, particularly its youthful population and linguistic diversity.

Speakers

**Speakers from the provided list:**

– **Ashana Kalemera** – Programmes Manager at CIPESA (Collaboration on International ICT Policy for East and Southern Africa), Session Moderator

– **Neema Iyer** – Founder and Executive Director of POLISI, a civil society organisation based in Uganda

– **Mlindi Mashologu** – Deputy Director General of the Department of Communications and Digital Technologies, South African government

– **Lacina Kone** – Director General and CEO of Smart Africa from Côte d’Ivoire

– **Matchiane Soueid Ahmed** – Special Envoy of the Mauritanian Ministry of Digital Transformation and Public Administration Modernization for the government of Mauritania

– **Shikoh Gitau** – CEO of KALA, a private sector company in Kenya (joined virtually)

– **Audience** – Various participants from the floor including representatives from Gambia, Ghana, and Liberia

**Additional speakers:**

– **Dennis Brumand** – Advisor of the Global Project on Digital Transformation, GIZ (mentioned as online moderator but did not speak in the transcript)

# AI Readiness in Africa: Building Sovereign and Resilient Digital Ecosystems

## Executive Summary

This comprehensive discussion on AI readiness in Africa took place at the Internet Governance Forum (IGF) in Oslo, hosted by the German Federal Ministry for Economic Cooperation and Development. The session brought together representatives from government, civil society, and private sector across multiple African countries to address how African nations can build sovereign and resilient AI ecosystems while avoiding digital neocolonialism.

The discussion revealed both the urgency and complexity of Africa’s AI governance challenge. With AI already impacting governance through automated disinformation campaigns and surveillance tools, while simultaneously offering transformative potential in agriculture, healthcare, and education, African stakeholders face the imperative of moving from passive optimism to deliberate intentional action. While 19 African countries have developed national AI strategies, the implementation gap remains substantial, requiring innovative approaches to financing, capacity building, and multi-stakeholder collaboration.

## Key Participants and Perspectives

The session was moderated by **Ashana Kalemera**, Programmes Manager at CIPESA, with **Dennis Brumand** serving as online moderator from GIZ. The discussion featured diverse perspectives from across the African continent.

**Neema Iyer**, Founder and Executive Director of POLISI, provided critical civil society insights, emphasizing that AI is already undermining governance in Africa through automated disinformation campaigns, erosion of public trust, manipulation of political discourse, and surveillance tools used to stifle voices at scale. She outlined civil society’s role as “watchdogs, advocates, educators, and storytellers” in AI governance.

**Mlindi Mashologu**, Deputy Director General of South Africa’s Department of Communications and Digital Technologies, offered governmental insights into developing national AI strategies. He emphasized transparent, inclusive AI governance frameworks grounded in constitutional values, while acknowledging significant infrastructure challenges including the digital divide and limited compute capabilities.

**Lacina Kone**, Director General and CEO of Smart Africa, provided a continental perspective that proved influential throughout the discussion. His assertion that “we cannot be passively optimistic, we have to be deliberately intentional” set the tone for the conversation. He highlighted that Africa is the number one frontier for data availability to train AI, yet over 1,000 African startups daily download APIs from foreign models, with training data leaving the continent permanently.

**Machane Babakar-Ahmed**, Special Envoy of Mauritania’s Ministry of Digital Transformation, contributed insights from a country actively developing its AI strategy. He emphasized that data sovereignty is as important as territorial sovereignty and that AI governance must be co-created with African voices, grounded in human rights with civic oversight.

**Shikoh Gitau**, CEO of KALA, participated virtually and brought private sector perspectives. Her pointed question about democratization—”Every time I hear democratising this, democratising that, I say, what’s the definition of democracy you’re talking about?”—prompted deeper reflection on whose interests are truly being served in AI policy development.

## Current State of AI in Africa: Challenges and Opportunities

### Existing AI Impact on Governance

The discussion began with Neema Iyer’s assessment of AI’s current impact on African governance, outlining how AI is already undermining democratic processes through automated disinformation campaigns, erosion of public trust, manipulation of political discourse, and surveillance tools deployed at scale. This established that the conversation was about governing AI systems already actively shaping political and social landscapes across the continent.

However, speakers also acknowledged AI’s positive applications in governance, including improved service delivery, enhanced agricultural productivity, and expanded access to healthcare and education.

### Infrastructure and Capacity Constraints

Significant infrastructure challenges constrain Africa’s AI readiness. Machane Babakar-Ahmed noted that in Mauritania, approximately 20% of the population in remote areas lacks adequate connectivity. Lacina Kone highlighted that Africa collectively lacks the necessary computing power across its 50+ countries to train AI models locally, creating dependency on external systems.

This infrastructure gap extends to human capacity, with speakers emphasizing the need for massive investment in digital literacy, technical expertise, and institutional capacity to support AI development.

### The Data Sovereignty Paradox

A central theme was the paradox of Africa’s data wealth and digital dependency. Lacina Kone observed that Africa generates abundant data for AI training, yet this data is processed through models located outside the continent. The irreversible nature of AI training—once data is used to train a model, it cannot be retrieved—adds urgency to sovereignty concerns.

## National AI Strategies and Continental Coordination

### Current Policy Landscape

The discussion revealed that 19 African countries have developed national AI strategies, representing significant progress in policy development. However, speakers consistently emphasized the gap between policy formulation and practical implementation. As Neema Iyer asked, “How do we operationalise beautiful policies and frameworks into actual ground-level implementation?”

South Africa’s approach, outlined by Mlindi Mashologu, emphasizes transparent, inclusive governance frameworks grounded in constitutional values. The country’s upcoming G20 presidency offers an opportunity to advance AI governance frameworks with focus on “solidarity, sustainability, and equality.”

### Smart Africa’s AI Council

Lacina Kone outlined Smart Africa’s AI Council for Africa, which focuses on five critical areas: computing power, datasets, algorithms, AI governance, and market development. The Council’s work on benchmarking the 19 national AI strategies represents an attempt to identify common elements while respecting national differences.

The discussion highlighted practical tools being developed, including the African AI Governance Toolkit (available at qbit.africa) and the African AI Maturity Index (at datawall.africa).

### Harmonization Versus Sovereignty

An audience member from Gambia expressed concern about countries working in silos rather than converging on a common continental AI policy framework. Lacina Kone’s response—that “not one size should fit all, but all sizes should fit together”—captured the nuanced approach required, recognizing that while harmonization is desirable, individual countries face different challenges and priorities.

## Multi-Stakeholder Collaboration

### Civil Society’s Role

Neema Iyer described civil society’s role as watchdogs monitoring AI deployment, advocates ensuring marginalized voices are heard, educators democratizing digital literacy, and storytellers documenting lived experiences of AI impacts. She emphasized the need for mandatory public interest impact assessments before AI system deployment.

### Private Sector Engagement

Shikoh Gitau highlighted the private sector’s role in creating enabling environments and conducting awareness campaigns. Her experience with teacher training campaigns across six African countries demonstrated high demand for AI education. She emphasized the need for blended financing approaches combining government resources with private sector investment.

### Government Responsibilities

Government representatives emphasized their role in creating enabling policy environments while acknowledging capacity constraints. Mlindi Mashologu highlighted the need for regulatory sandboxes and startup support mechanisms to encourage AI-enabled economic transformation.

## Data Sovereignty and Cultural Preservation

Machane Babakar-Ahmed’s assertion that “data must remain in African hands” and that “digital sovereignty is as important as territorial sovereignty” provided a framework for understanding sovereignty in the AI era.

Lacina Kone emphasized preserving Africa’s 2,000+ languages through locally trained AI systems, highlighting the cultural dimensions of AI sovereignty. The potential for AI to reach indigenous people in rural areas and educate them in their own languages represents a transformative opportunity.

## Economic Development and Innovation

### Redefining Success Metrics

Lacina Kone’s observation that “Africa is not looking for the most powerful AI, it’s looking for the most useful one” represented a fundamental reframing of AI development priorities. Rather than competing on computational power, this approach focuses on practical applications addressing African development needs.

### Concrete Examples

The discussion included specific examples of African AI initiatives, such as the African scientific panel with 25 doctors of African heritage working on healthcare AI benchmarking, demonstrating practical approaches to developing contextually relevant AI systems.

## Audience Engagement and Concerns

The lightning round featured questions from multiple African countries. A participant from Ghana raised concerns about youth fears regarding AI taking over jobs, highlighting the need for trust-building and proper education. A representative from Liberia asked about realistic steps for implementation, emphasizing the need for practical guidance beyond policy frameworks.

## Education and Capacity Building

A consistent theme was the need to democratize digital literacy and make AI education accessible at grassroots levels. Neema Iyer observed that elite policy discussions often don’t apply to most people living on the continent, highlighting the need for accessible communication strategies.

Shikoh Gitau’s teacher training campaigns provide a model for building capacity and confidence simultaneously, with high demand for AI education when delivered appropriately.

## Key Challenges and Future Directions

### Implementation Gap

Despite progress in policy development, the question of how to operationalize frameworks into ground-level implementation remains central. This reflects broader challenges including limited resources, capacity constraints, and coordination difficulties.

### Infrastructure and Financing

Fundamental infrastructure challenges—including limited computing power, inadequate connectivity, and capacity gaps—require innovative financing approaches. The scale of investment required compared to available resources suggests difficult prioritization decisions ahead.

### Youth Engagement

Concerns about youth fears regarding AI and employment displacement highlight the need for effective communication strategies that build understanding of AI’s potential benefits while addressing legitimate concerns about economic disruption.

## Conclusion

The discussion demonstrated both the complexity and potential of Africa’s AI governance challenge. The emphasis on moving from passive optimism to deliberate intentional action captures the urgency of the moment. Africa’s abundant data resources and youthful population provide significant advantages, but realizing this potential requires coordinated action across policy development, infrastructure investment, capacity building, and institutional innovation.

The conversation revealed a mature understanding of Africa’s position in the global AI landscape, with stakeholders articulating a distinctive vision that prioritizes utility over power, inclusion over efficiency, and sovereignty over dependency. The path forward requires sustained commitment from all stakeholders, innovative financing mechanisms, and continued collaboration across national boundaries while addressing substantial practical challenges.

As Lacina Kone emphasized, the goal is not to develop the most powerful AI systems, but the most useful ones for African contexts, suggesting a sustainable development pathway that builds on existing strengths while serving African needs and values.

Ashana Kalemera: Music Good afternoon. Thank you so much for joining us this afternoon. I’ll also say good morning, good evening and good day, considering that there are participants joining us online from different time zones. Welcome to the session on AI readiness in Africa in a shifting geopolitical landscape. The session is hosted by the German Federal Ministry for Economic Cooperation and Development, BMZ, together with its partners on stage. I’m very, very, very honoured to be moderating this very timely discussion. My name is Ashna Kalemela. I work as programmes manager at CIPESA. CIPESA is the Collaboration on International ICT Policy for East and Southern Africa. I’m sure we’ll all agree with me that there’s huge transformative potential of AI in society, from innovation to socio-economic development. However, there are also very significant risks. These include inadequate governance frameworks, which risk deepening inequality, weakening democracy and reinforcing technological dependencies. In Africa particularly, countries are striving to build sovereign and resilient AI ecosystems in tandem with a fast-evolving geopolitical landscape. This landscape is marked by shifting alliances, intensifying technological rivalry and growing economic pressure. Whereas various stakeholders are engaging on these issues, the continent remains underrepresented in global AI development. as well as discourse. Locally driven solutions are constrained by limited investments in research, regulatory gaps, and the dominance of multinational tech companies. Meanwhile, concerns about digital exploitation and economic disparities when it comes to data processing, training of models, and low-wage labor markets in Africa also prevail. The risk of AI-driven digital neocolonialism is growing. As global powers compete for technological influence, Africa must strengthen its position to ensure AI serves local needs rather than external interests. African nations at the moment have a unique opportunity to establish AI governance models that are rooted in fairness, in transparency, and inclusion. These AI frameworks also have the potential to align with local realities and normative considerations. These frameworks are hopefully able to foster innovation, uphold democracy, and human rights. Our speakers today, who represent a very broad spectrum of stakeholder groups, will highlight the challenges and opportunities for securing an AI future that benefits Africa. The speaker lineup includes Shikoh Gitau, who’s joining us virtually. She’s the CEO of KALA, a private sector in Kenya. We have Mr. Lassina Kone, the Director General and CEO of Smart Africa from Cote d’Ivoire, Enchanté. Machane Babakar-Ahmed, the Special Envoy of the Mauritanian Ministry of Digital Transformation and Public Administration Modernization for the government of Mauritania. We have Melindi Msalango, the Deputy Director General of the Department of Communications and Digital Technologies, the South African government. And to my immediate left, a very old and good friend, Nima Ea. the Founder and Executive Director of POLISI, a civil society organisation based in my home country, Uganda. Our online moderation is being done by Dennis Brumand, the Advisor of the Global Project on Digital Transformation, GIZ. You’re all welcome once again, and we’ll kick off the conversation with a lightning round, which is a one-sentence response I expect from the speakers here. And the question is, what must be done today to ensure that AI strengthens rather than undermines democratic governance in Africa? I’ll start with my left over to you, Nima. Thank you so much for the question, and I’m very pleased to be here. I’ll start again. Hi, everyone. I’m very pleased to be here. Apologies for the mic malfunction.

Neema Iyer: So, I was saying that AI is already undermining governance in Africa, and we’re seeing this through automated disinformation campaigns, the eroding of public trust, manipulation of political discourse, and surveillance tools that are used to stifle voices at scale. These harms are often gendered, and a lot of our work is looking at a feminist perspective on new technological tools. And these harms are magnified in contexts that have weak data protection and limited digital literacy, which really opens the door to abuse. So, as such, I would say that we need AI that is based on accountability, trust, and a big component of public education. And this will include, for example, impact assessments, regional coalitions, investment in ethical and open-source alternatives that work within our context. and are based on our realities and that are focused on care and justice rather than control and extraction. Thank you, Ashna. Back to you. Thanks, Nimba. Same question to you, Mr. Malindi.

Ashana Kalemera: What must be done today to ensure that AI strengthens rather than undermines democratic governance on the continent? No, thank you. Thank you, Margaret. I think on my side, I just would like to say that what is important is that we need to institutionalize the transparent inclusive AI governance frameworks that are grounded in the constitutional values of public participation to ensure that AI supports accountable service delivery, social justice, as well as democratic resilience. But also some of the areas that I would like to highlight is the issue of removing bias from the data sets that are used to train the AI systems.

Mlindi Mashologu: Because if we are not removing these biases and also not including, you know, a large pool of, you know, demographics in terms of data sets, you’ll find that you can have, you know, the challenges, you know, and which will then undermine, you know, the democratic governance. But also, I think the last one that I just want to highlight is the explainable AI, whereby we would like to advocate to say that whatever decisions that are taken by the AI system, they need to be explainable and they need to be based on, you know, some human oversight, you know. So I think those are some of the areas that I would just like to highlight on my side. Thank you. Thank you very much.

Ashana Kalemera: Moving to my right, Mr. Kone. Thank you very much for inviting me. This is my application in honour of, of course, our partner, BMZ, In fact, can you hear me?

Lacina Kone: What you just mentioned, I would say the impact is actually worse than all of this. We talked about the bias of the AI. Don’t forget, today even if you have a PhD degree, you can still be ignorant in terms of AI. If we go based on a foundation that is not enough to go to school, but you have to adapt to AI, and we all know that the inclusions definitions in Africa is completely different from the inclusion definitions in the West. Because we believe that the AI will be the equalizer, be able to include people who do not speak any other language. Remember, we have more than 2,000 languages in Africa. So if we allow those languages to be trained on the AI system, but not trained by us, it basically means even indigenous people will be impacted by any cultural biases using AI. Therefore, all what my predecessors have just mentioned is very true. So what do we do about it? We cannot be passively, we can no longer be passively optimistic. We have to be deliberately intentional. That’s why Smart Africa put together last April Africa Council of AI Council for Africa. What does that actually mean? The Africa AI Council for Africa, the AI Council for Africa will look at five different things. One, computing power. Collectively today, more than 50 countries on our continent, do we have a necessary power and a computing power to be able to train our data. Two, it’s going to look at the data set. No matter what other people said, Africa today is the number one frontier in terms of availability of data. to train AI. Number three is going to look at the algorithm, algorithm which is actually the cultural bias, what goes on into the AI to be able to respond to people. And number four is the AI governance. In the AI governance side today, there are today about 19 countries in Africa who have already developed their national AI strategy. It’s our role at Smart Africa to be able to harmonize those policies, to be aligned exactly to be what the high level of the UN commissioners, they came up together with the AI governance, as well as looking into what European have put together in terms of safeguarding. Of course, all of these have to be aligned with the AI strategy developed by the African Union. And number five is the market. I think what is happening today, what has happened, happened. But what do we need to do about it? We need to be doing these things intentionally. It’s actually worse than you actually want to think. Because currently in Africa, on our continent, there are over a thousand startup who are downloading on a daily basis, the API from Open Frontier, Open AI Frontier model, as well as the DeepSeek, which is the Chinese. But they’re training those model. Those models are not located on our continent. They’re located outside the continent. And AI system is the way once you train the server, you can never get the information back again. So what do we do about it? It actually goes to even to look at the foundational model of our relationship with our partners in Europe. What do we do? Based on the fact that if you look at North America, everything is based on the private sector base, which is at the heart of the capitalism. When we look at this, it’s based on the control of the government. And Africa wants to go with the user-centric approach for everything. So it’s time for us, we actually be able to look at the AI, not the most powerful one, but the most useful one, by looking at our values, by looking or preserving our languages. to be able to create opportunity, not for a few, but for the many. Thank you.

Ashana Kalemera: Thank you very much, Mr. Kone. Over to you, Madam Ahmed.

Matchiane Soueid Ahmed: Thank you very much. Thank you very much, moderator. I’m fully agree with what my colleagues have mentioned right now, but I would like to summarize it in two words. First, to ensure AI makes democratic governance stronger in Africa, we must invest now in inclusive, transparent AI policy frameworks that are co-created with African voices. Second, these frameworks must be grounded in human rights and supported by strong civic oversight. Thank you very much. Thank you all for the very quick lightning round. We’ll now deep dive into specific questions, per sector, per experience, per expertise. And I’ll stay with you, Madam Ahmed. Mauritania was amongst the first African countries to launch a national AI strategy. What key challenges have you faced in translating this strategy into tangible actions? And how are you addressing coordinated institutional ownership and inclusivity to ensure a democratic and locally rooted AI in Africa? Thank you very much for this important question. Yes, Mauritania has developed an AI strategy with the support of German cooperation through BMZ and GIZ. But to implement that strategy, we are facing some challenges. Let me first start by infrastructure availability in rural areas. As you may know, Mauritania has a huge surface, about one square kilometer. And there is a lot of small villages. far away from each other. That makes it difficult to serve more than 20% of the population living in this area. So we are facing this reality, but we hope to address it as soon as possible with the support of government partners, like German cooperation through KFW, and European Union too, and World Bank. Another challenge is limited capacity and technical expertise, which slowed this strategy implementation. To address this, we are investing in partnerships with universities and international organizations to build local skills and knowledge. We faced also a coordination problem across government and sector, and to tackle this, we established an inter-ministerial working group to ensure alignment between the AI strategy and national development priorities with regular consultation among stakeholders. We are also aware that ownership and inclusivity are crucial. That’s why our approach has focused on creating a participatory process, engaging civil society, local tech communities, and youth in shaping policies and pilot projects. This will ensure that our AI applications are relevant, trusted, and anchored in local needs. Ultimately, our goal is to build an AI model made in Africa that reflects our values and development priorities, inclusive, ethical, and locally learned.

Ashana Kalemera: Thank you very much. Thank you very much, Ms. Ahmed. I’ll come back to you. Mr. Kone, having had the government perspective and being an intergovernmental organization, from your experience, how can Smart Africa help its member states build AIR governance models that protect national sovereignty while also advancing democratic values and public interest, as Ms. Ahmed has pointed out?

Lacina Kone: Thank you very much for the questions. In the current geopolitical situations, and I like to see this all the time, the multi-stakeholder multilateralism is a choice, but we’re living in a multi-polarity geopolitical moment. The only thing that should bind us together is our differences. That’s why Smart Africa is here. It means we are not looking for a national AI policy, it’s like a cookie cutter. One side should fit all, but all sides should fit together by preserving a human right and a digital right. And this is all drawn from the UN high level. If you look at the AI governance of the UN high level, they are structural. They are basically regulatory issue, they are national issue. The national AI governance may be a little bit different, but if you look at the common denominator based on ethical inclusion and sustainability, it has to be included in a national strategy. However, by saying that today Africa, no single nation will be able to build an AI system alone. My sister just mentioned about infrastructure. But you see, we need to understand, some people said, oh, Africa, you might be so behind because you only have a 40% of your population covered with the internet use, then why are you talking about AI? That’s not the question. Africa is not looking for the most powerful AI, it’s looking for the most useful one, looking at the agriculture, looking at the healthcare and looking at the education. Why? Because if you know the history, for the past 45 years since 1980, African continent has gained 1 billion people in 50. But in 45 years, if you gain 1 billion, has the number of school increased to that rate? No. Has the hospital increased to that rate? No. Has the school and the financial sector bank have grown to that rate? No. It means for us, digital transformation, when we look at the AI, it is the evolution within the ecosystem of digital economy. But it’s a revolution within itself, which means AI allows us today to actually reach indigenous people in the rural area, to educate them in their own language, so they become tech-savvy like anyone who’s been to the university. So the way we are doing that, Smart Africa, we do have a Council of ICT Minister, where Mauritania is a part of it. We have a Council of African Regulator. We have a Council of African IT Agency. And we have also the board member who are the head of the CID himself. That’s why we created AI Council, to be able to address all of this methodically and systematically in a way that we align and we mutualize our resources to be able to face this revolution, which is the AI. Thank you. Thank you. I don’t know about the audience, but I’m here nodding and taking notes frantically. We had hoped to have private sector perspective complement the government and intergovernmental ones, but unfortunately, our virtual speaker, Shiko, has not been able to join us. Oh, sure. Please go ahead, before I go to civil society. When we talk about a private sector perspective, I want ladies and gentlemen, all of you to remember, today when any nation talks about AI, they talk about a frontier model. We’ve seen the American talking about a frontier model, open AI, LAMA, GORC3, what name of. When we hear the French people talking, they are behind the mistrial. We hear about the Chinese talking about the AI, they’re talking about the deep-seek. All of these people are private sector, which basically means the government in Africa knows very well their role. They should be creating a conducive environment for private sector to have a kind of blended financing. Because when we talk about financing, America will declare about 500 billion, French about 200 billion. So what is Africa going to declare in terms of funding? This fund will be and should be a blended. Because don’t forget, we have over half a dozen MNO, which is a mobile network operator operating on our continent who owns data center. It is existential for them to be able to get into the AI because at the end of the day,

Ashana Kalemera: it’s about making money. So even in our financing, we should be taking that into account. That’s where the role of the private sector comes in. Thank you. Thank you very much, Mr. Kone. Shiko has joined us online, but before we move to him, we’ll first hear from civil society and then additional perspective from government. So Nima, civil society is crucial for ensuring that AI governance serves the public good. What unique contributions can civil society make in this space and what kinds of frameworks or coalitions are needed to support effective oversight and inclusive participation? Thanks so much for the question.

Neema Iyer: So of course, civil society is extremely crucial to the entire process. And I’m not just biased in saying that, but we play a role as watchdogs, as advocates, as educators, as storytellers, and shaping the entire narrative about AI on the continent. And we’re also there to question the political, the economic, and the social logic on why we’re actually deploying AI. Shikoh Gitau, Dennis Mwighusa, Shikoh Gitau, Dennis Mwighusa, Shikoh Gitau, Dennis Mwighusa, We need to resist this quiet import of oftentimes harmful tools, you know, such as excessive surveillance or predictive tech. There’s so much good that can come from AI, but of course there’s also a lot of harms and I think we should have a very balanced view of both. So we can’t just be on the side of, you know, AI is harmful or AI is beneficial, but really having that balanced view to think critically about why we’re bringing AI in. The second one that’s really important is to track the funding. to understand who is funding what in the African context and what are these different foreign interests of all these players, of all the tech giants in shaping our AI agenda. I think we need to be very critical about that as well. The third one I would say is that we need to document the lived experiences, both of the benefits and the harms, especially to marginalized communities. And, yeah, we can do tech audits, of course, but I think civil society is uniquely placed to address these harms and benefits and to document them and to tell these stories so that they then go back and inform the policies. The fourth one I would say is we really need to democratize digital literacy. I feel like, you know, we’re in Oslo here having this conversation that doesn’t apply to most people living on the continent. I think there’s really a need to take these conversations to a grassroots level. It is extremely important for us to be represented at these high-level multi-stakeholder meetings, but that doesn’t mean we do it at the cost of not involving local communities at different levels in this conversation. And, yeah, I think the education needs to be accessible. It needs to be forward-looking. It needs to be for common people, for leaders, in local languages. We need to use creative ways of talking about it. We can’t always use these huge data governance languages, and then, you know, people’s eyes glaze over because they don’t really know what you’re saying. We really need to make it accessible. I think that that is a very urgent need. The fifth one I would say is that we also need to get back to making things. And, yes, I agree with the point. We do not have the investment of that value, but it doesn’t mean that we don’t do anything. I want to propose that we have alternative design of products, of AI products that are based on different incentives and that, you know, are really tailored to our needs. And so, moving to the question of, like, what kinds of frameworks I would say that we should really have mandatory public interest impact assessments, so I touched upon that earlier. But it would be lovely if before an AI system is deployed that, you know, we do assessments of what it can impact, who it will impact, and that, you know, throughout the process of having it, we continue to do these assessments. And we understand how it impacts, you know, social, economic, gendered, environmental impacts of these AI systems. The next one, I think we need regional civic coalitions, so it’s amazing to bring governments together to talk about AI, but what other groups are we leaving out? What silos are we creating when we bring these groups together? Who is missing at this conversation, at this table, for example? So, you know, women’s rights organizations, trade unions, journalists, researchers, can we all come together to shape these policies and that they’re not just done at a very high government level? The third one, again, going to my colleague’s point, is funding. We really need better frameworks for how we are going to fund AI on the continent. We need to come up with our own funding models for innovation in a way that is sustainable, in a way that we bring young people in to actually create product. I feel like that is quite a bit of a missing link. And then the last one, I think, is, yeah, just bringing as many stakeholders into the conversation as possible, and then just questioning how are we going to operationalize all these beautiful policies and frameworks that we’re talking about? So how do we take it from this table here, actually putting it on the ground and actually seeing it in action? So thank you so much. Thanks, Neema. What you’re saying resonates a lot with me, coming from civil society, but absolutely what was also said from the government and intergovernmental perspective. Malindi, I’ll move over to you. In anticipation of South Africa’s G20 presidency, How can AI maturity assessments be leveraged to strengthen context-specific governance across Africa? What role do institutional readiness, democratic safeguards and inclusive policy-making play? Thank you for that question.

Mlindi Mashologu: As the country, South Africa, we assume the G20 presidency and I think it’s important to note our banner there is solidarity, sustainability and equality. So we see that we are presented with a unique opportunity to lead a new chapter in the digital governance, not only for ourselves but also, you know, for the community at large. I think it’s important to note that, you know, one of the powerful tools that we can use is the AI maturity assessments. That is one that we did work with the GIZ in terms of South Africa, which we participated on and one of the areas that we picked up is that, you know, these assessments, they allow us as governments to diagnose the strengths, you know, identify the gaps, but also chart, you know, clear actionable pathways for the responsible AI adoption. But also it’s important to note that they are not just technical diagnostics, they are also political governance instruments. So if we apply them with PIPOs, you’ll find that they can actually be transformative in nature and I think they can actually assist in terms of the African-led, you know, governance frameworks. But now if I can just also look into, you know, how we can anchor them into the African realities, you’ll find that too often, you know, the benchmarks are normally imported wholesale benchmarks, which are normally failing to account for. and other governance challenges such as limited compute infrastructure, fragmented data ecosystems or linguistic diversity. And these are some of the area things that are very critical when you look into the African continent. So for then the South African and our continental peers, so we are advocating that we need to develop and localize these tools to reflect our developmental priorities, which then include inclusive service delivery, ethical public sector automation as well as community trust. But also it’s important to recognize that institutional readiness is the cornerstone for any AI governance framework. We do understand that as policy makers, policy alone are sometimes not enough. But we need some of the institutions that are technically equipped, policy-coherent and operationally agile. So one of the areas that we are currently doing as a country is the development of an AI policy for the country. And I mean the policy, you know, it does have broad statements that we are looking into. And I must say that we have been quite behind in terms of that. But at least we are on the final stages. And I think if I can just highlight just a few areas that we are looking on as a country. One is the area of capacity development, where we are looking at strengthening the AI-related education. And these are some of the things that also came, you know, from when we are unpacking, you know, the policy, you know, the frameworks that we looked into. But also we’re looking on the areas of AI for economic transformation, where we are looking in terms of AI as well in public service delivery, but also supporting the startups as well as through regulatory sandboxes. The other area we’re looking on is the area of responsible governance. So we are advocating on some of, you know, various bodies that needs to be established, which includes your ethics board, AI ethics board, national AI. the AI Commission and the AI Regulatory Authority because we feel that some of the regulators that we’ve got might not be, you know, up to people in terms of, you know, regulating AI, but also the areas of ethical and inclusive AI, where we’re looking at developing, you know, localizing ethical standards, but also cultural preservation and international integration, as well as human-centered approach. And, I mean, if you were to look on all these broad aspects, you’ll find that, you know, some of the work that the colleagues are saying, it also aligns to that, because you’ll find that, you know, while we’re developing these policies, you’ll find that there are some challenges that are deep-rooted into our societies, which include now currently, if you look on digital divide, you’ll find that we still have got digital divide as a continent, which we need to address to make sure that, I mean, whenever we put, you know, these frameworks, they can address that. But also, if you were to look on the compute capabilities, you’ll find that, you know, we still don’t have compute capabilities in the continent. So it’s one of the things, then, that we need to make sure that we address significantly as part of us developing the governance frameworks. But I also want to add the other area, that we also need to embed, you know, the democratic safeguards into every layer of the AI development and deployment, because we see that without the robust mechanisms for transparency, public accountability, as well as end recourse, AI can deepen exclusion, entrench bias, or even erode civil liberties. So a democratic governance in this age of AI means placing human rights and constitutional values at the centre, from the procurement processes to the algorithmic audits. And I think the last point that I want to add critically is the role of inclusive policymaking. AI governance cannot just be a technocratic rule, but it needs to be

Ashana Kalemera: Thank you very much, Melendi. We are looking very keenly and proudly at South Africa as it steers the G20 presidency and we hope these are issues that will be driven forward during the tenure. Our fifth speaker online, Shikoh Gitau, was able to join us and I would like to put the question on private sectors role back to him. Private companies are at the forefront of AI innovation, Shikoh. From your experience, what governance frameworks are necessary to ensure this innovation aligns with democratic values and how can the private sector contribute to trust, transparency and accountability of AI, which have been resounding issues from all the speakers here? Thank you so much for having me. Can you hear me? Yes. Thumbs up. Awesome, thank you. So, I’m joining in from very cold and rainy Nairobi,

Shikoh Gitau: and I’m really glad to be here. Thank you so much for having me. And apologies for joining in late. So, the question around private sector and what private sector can do is start with building an enabling environment. A couple of months ago, I was speaking to a group of, like, policymakers and government asking, what do you need? We just need governments to, A, understand the potential of the space we are in. AI is not a passing fad. It is not a buzzword that is going away. It’s a consequential technology for our generation. And we need to be able to take care of that. And by building an enabling environment, I was hearing the interventions from Engineer Kone, and those ones from the representative from South Africa, and they’re quite interesting. But how do you then move that from paper to pavement? How do we do that? Is incentivizing private sector to come? And I hear the conversation around compute, around talent development. What does that actually mean in actual sense? I’ll give an example is, we have been working with a number of, like, A, governments on one hand, but also startups. And trying to understand is, what do you need for you to succeed? Because there’s so many pilots that are happening on the African continent, but they’re not getting any traction. But then you soon realize is that we are building all these things in the AI space, but our market does not understand AI, does not know what AI is, and the potential that AI has to. And what’s the difference between what we’re doing right now and what they were being sold for five years ago with mobile internet? And being able to concretize that to our Our market base, our customer base, our user base is critical. And while it is the work of the government to educate the populace, it is the work of the private sector to actually create massive awareness. But if the government comes in to interrupt this, there is a challenge. So for example, this week, we are running an AI awareness and fluency and literacy campaign across the continent in six markets, in six countries, and what we are targeting is teachers. And you’re not going through the normal, going through ministries of ICT or ministries of education. We are working with teachers directly through their associations, through their communities to train them. And in this, we said it last week, and in this week, what the difference and the demand is very different. The first one, we did a baseline study, there’s a lot of fear around AI. What we are hearing right now is, can we continuously do this training over the next six months? What is the incentive that the governments that we are working with have done is they’ve given us free space to create this programming. And they said, once you’re done with the pilot, because we are calling it a pilot, send us a report. That is a very concrete way of showing private sector is we as government don’t know what actually needs to be done, but if you can pilot and give us results, you can be able to scale whatever you’re doing, or we can be able to create even more space for you. Similar to investment in the startup sector in compute. So enabling private sector to invest in compute does not mean hindering what government can do in compute. I agree with Engineer Connors saying that funding for Africa has to be a blended instrument. And it means that private sector will have to invest in some of these compute facilities. be extremely expensive. But on the other hand, government and donor organizations have to invest in the earliest stages of this compete because we need to be able to have researchers and startups building and they cannot be able to afford the enterprise-grade computing that will be sold to them by enterprise versions. So it is bringing this blended thinking and creating the enabling environment. From a democratic point of view, every time I hear democratizing this, democratizing that, I say, what’s the definition of democracy you’re talking about? And for me, democratizing means that it’s enabling everybody everywhere to have access to the same opportunities and resources. And if you’re going to be able to bring that democratic tenancy in AI, it means that even the policies that we are making are not being prescribed to Africa. So what we have seen is we need policies and policy frameworks, as rightly said by South Africa, that are enabling the ecosystem. But who is drafting these policies? What agenda do they have? Do they have Africa at heart when they are doing this? Those are the questions you should be asking. And who is doing this? It’s enabling young Africans to be able to contribute to some of this conversation. So, for example, one of these things we are doing with, we actually support from Smart Africa, is this African scientific panel that is calling to young Africans, both in the continent and in diaspora, to enable and support their governments and governments across the continent in drafting some of these conversations, some of these frameworks that you’ve spoken about. But beyond drafting and writing them out is bringing them to life. So, for example, there’s a compute project and a benchmarking project that is being run by 25 doctors from across the world with African heritage to look at. Are these models for health care responding to African needs and African reality? That is what democratizing means. It’s creating, again, it starts with creating the enabling environment, getting out of the way, but also then starting to resource and accepting health from the democratized environment.

Ashana Kalemera: Thank you so much. Thank you very much, Shikoh. The clock, I tell you, is moving much faster than it should, so we need to speed this up a bit. Because we’re here to learn and exchange, I’d like to give an opportunity to our participants to share any comments, reactions, or questions to the five speakers. There are microphones on either side of the room there. Please feel free to walk over, introduce yourself, and share your comment, reaction, or question. As those in the room make their way to the microphone, I will also check with our online moderator if there are any questions from the online participants. No questions from the online participants, but I see one participant walking to the microphone. The floor is yours, sir. Okay. Good afternoon, and thank you very much. I’m a little bit tall. Good afternoon, everybody.

Audience: My name is from the Gambia, and the question actually I have is with all, it seems like each African country, or most of them, are on the path of creating their own policy. And I’m trying to find out from Smart Africa in particular, what can you do to bring all these countries on the table where I know that each country has your own, you know, different way of doing things, but at least there must be a convergence where we can actually have a policy that could be used from all across the country, like we had in the African Union Convention on the Cyber Security and Personal Data Protection. Can we have something that, you know, that goes across the continent instead of everybody being in silos, and again, it’s going to take a very long time. before that is done. So I want to ask Mr. Kone to see if they can bring everybody on the table, at least try to work onto a common platform. Thank you.

Ashana Kalemera: Thank you very much, Alhaji. We have a second question. We’ll take that before coming back to the speakers.

Audience: Good evening, everyone. Is it? Okay. My name is Lydia Lamisa Akamvareba from Ghana. I’m looking at the team up there. AI readiness in Africa in a shifting geopolitical landscape. Yes, it’s true. But we need those kind of policies that will create the enabled environment that will bring trust in the African society. Because currently, Africa is one of the youngest population that we have in the world. And the youth has a challenge, whether AI is coming to take over their jobs. So we as policymakers need to build that trust. I want to find out what are we doing to bring that trust among the youth in Africa? Thank you.

Ashana Kalemera: Thank you very much, Honorable Lydia. Yes, sir. If you walk a little bit quicker to the microphone, we’ll take your question. Good day, everyone. My name is Peter King. I’m from Liberia. Mine is in the form of a comment as well. I realize that there are four key areas that I can

Audience: talk about. In Africa now, what are we looking at? A lot of people work on policies. But we’re forgetting to look at the issue of digital literacy, the level of infrastructure. And these are things that once we are looking at readiness, this should come to the fore. So to the panel up there, thank you so much for the nice presentation. Can you help us to understand better, from our context from Africa, what are the realistic steps that governments can do? to ensure that this is a reality in terms of being ready, to ensure that youth alone is claimed and is real that Africa has a youthful population. These are challenges. We’re looking at what are the opportunities. Can they be leveraged for the youth to use it? We are always looking at, we are always using it, the OpenAI, the ChargeGBT is not created by Africa. So where do we go from here?

Ashana Kalemera: Thank you so much. Thank you very much, Peter King. So we have 11 minutes to the end of the session. So we’re going to do a crash course in closing remarks, responses to questions and final thoughts, all in one go. So I’ll invite Mr. Kone to take the question that was directed specifically at you. And then for the rest of the speakers, we’ll reflect on the questions and issues around trust and realistic steps that have come from the floor, while at the same time, sharing concluding remarks on what the future should look like five years from now. So Mr. Kone first.

Lacina Kone: Thank you very much for the questions regarding policy harmonization on the national AI strategy. As I said before, what really matters is not like one size should fit all, because each country has a sovereignty you have to take into account. But we have to make sure that each size should fit together. So what we’ve engaged in since April, we’ve actually scanned our continent. We found there’s close to about 17 to 18 countries who already develop, including, of course, Nigeria, including Mauritania, who have developed already the national AI strategy. So what we did, actually, we took all of these national AI strategy, and we also looked into the AI governance regulatory environment developed by the high level of the United Nations. What it entails exactly and we actually look also into the African Union AI strategy So we try to do a comparison to come up with a benchmarking What are the countries saying are they addressing exactly what is required from the United Nations? How does it compare to Europe and how does it compare also to the African Union? so we’re coming out with the benchmarking and this is also part of the work we are doing for the AI Council for Africa Starting from this July, of course, we’ll be sharing we’ll be having a CMICT Which is a council of ICT ministers and we’ll be sharing with them We’ll said if you look at the national AI strategy of Nigeria What was the best lesson learned from that compared to what Egypt has? Because you might see that Egypt for example is probably skewed towards more startup Action the startup hubs and creations where you while you will see that Nigeria is actually focused on something else So we’ve compiled all of these and of course, we not only talking we actually walking or talk what I’m saying Exactly. This would be shared with our Council of ICT minister which from there it goes to the regulator But it’s very true so far now we have about 19 country who have a different AI strategy But we need to remember something if you look at the fundamentals which are the common denominator looking at the ethical use of the AI looking at the inclusion of the AI inclusion of the Inclusivity if I said and look at the sustainability 90% of those a national strategy it address it. However, you may Elaborate the best AI national strategy, but is it conducive for private sector investment? That is the one question we need to ask because at the end of the day Government should be creating law and regulation to create a conducive environment for private sector to be able to chip in. Thank you

Matchiane Soueid Ahmed: Thank you very much. Mr Still on my right Ms. Ahmed, reactions to the questions from the floor, as well as concluding remarks on what the future of AI and the continent should look like. Thank you very much. I want to comment some reaction of the audience. I totally agree with that, working in SILO will not solve the problem, at least we need to look after a synergy between different initiatives. But the important question is how, because as you know, the context of each country is different, and even the context of the region in Africa is different. So we need to think out of the box to find solutions that are suitable for everyone’s context. So about the future of Africa AI, as we shape African AI, we must lead with principles. Data must remain, in my understanding, data must remain in African hands. Because as you know, in the era of digital and especially in the era of AI, sovereignty is not just territorial, but also digital. Second conclusion that I can mention, that equitable and democratic AI is only possible if policies, power, local communities, protect citizen rights and foster homegrown innovation. Thank you very much. Thank you very much Machane. I’ll move to my middle now, because I see Chico on my screen in front of me.

Ashana Kalemera: Chico, reactions to the questions from the floor, as well as concluding remarks on AI’s future on the continent. Thank you very much. I want to respond to the question that the gentleman from The Gambia

Shikoh Gitau: asked. So for the past year and a half, we were looking and doing a scan on what is the right governance framework for Africa. And what we concluded is not everybody needs a strategy. A strategy needs to be thought through. But what are the instruments that need to be in place for you as a government to be able to start on this AI journey? And so what we have developed and what was launched in April was the African AI Governance Toolkit. If you go to qbit.africa, qbit.africa, you should be able to see it. And it enables you as a government official, a regulator, a ministry to be able to see, should I just have like AI principles that are guiding our work around AI? Should we have a policy framework? Should we have a strategy and a simple cheat sheet that you can walk through and work through to help you develop the principles? I mean, as outlined by Engineer Kone of safety ethics and such to be able to develop like a very simple policy framework or principle guideline or a strategy that will help you on this AI journey, because we recognize that not every country is able to invest in having an AI strategy. It’s a little bit premature. There are a couple of countries you’re working with in Central Africa where they’re starting with a digital strategy, not even an AI strategy, but they need to start working on AI. So a simple framework is what we are trying to work with them to be able to start on the journey. On the second one, I love the second question because from Ghana, and it’s a very critical question, but I always say you cannot… work on what you cannot measure? So we’re talking about talent, a maturity, all these questions. Does anyone know where you stand? Does Ghana know where they stand in terms of AI maturity on the African continent or in the world or where you stand about your talent readiness? So again, the other instrument that we’ve worked, and again, as Minister Kone, I mean, Engineer Kone mentioned around the AI Council is developing the African AI Maturity Index, which is a live document and a live tool, not a document, a live tool that is constantly being updated as the different strategies are being put in place, different policies are being put in place. So if you go to data wall, d-a-t-a wall, w-a-l-l.africa, you’ll be able to see the maturity. We’ve broken it down to very, very simple metrics, broken down to the schools that are teaching STEM in our country. So it’s very, very granular to help you realize, as you’re solving for this AI, maybe we should start introducing more STEM courses, not even AI courses in our country. The second one is the Talent Index, which is primarily focused- I’m going to interrupt you. I’m finishing. I’m finishing. This is my last one. All right. Go ahead. This is measuring what talent actually means and what you need to put in place. Again, it’s being very actionable. What should Africa do? I’m very bullish about Africa. If you’ve heard me speak, is I always say Africa can, and we can win this AI race because, as I always quote Minister LACINA KONE is, because we are defining what the race is about. We are not following other people’s rules. Thank you very much. Thank you, Shikho. Somelindi, and then Nima.

Ashana Kalemera: The race is on. One minute each. No, thank you. I think I just want to touch a bit on what the other questions were.

Neema Iyer: where I’m doing my PhD. A lot of students are submitting AI assessments. What are you grading as a teacher? So the whole way that we even assess skills needs to completely change. And in terms of my closing remarks I think we need to continue to have radical and critical questions about AI. Always, always, always question everything. What are the harms? What are the benefits? How is it serving our future? How is it serving the future of our youth? Thank you so much for having me. Thank you. 28 seconds for me to wrap up this very rich, rich discussions. Quite clearly, it’s not yet Uhuru for AI in Africa, but given what’s been discussed here with the right investments the continent stands to benefit significantly. I think the areas for prioritizations that have come through include pushing for accountability, public education, transparency, inclusion, not only from a language perspective, a gender perspective, and ensuring co-creation, infrastructure investments, multi-stakeholder consultation and participation, impact assessments, that was a very, very good one, funding, procurement, and ensuring that, you know, policies are not just prescriptions, but actually live up to reality. Thank you so much.

Ashana Kalemera: We appreciate the perspectives that have been shared from Côte d’Ivoire, from Mauritania, from Ghana, from the floor, from Gambia, from South Africa, as well as Uganda. Once again, thank you to the speakers. Thank you to the organizers, BMC. We appreciate the conversation and ensuring that discourse about AI readiness on the continent is making it to the global stage, i.e. the IGF. Thank you very much. We wish you all a very good evening, and thank you for attending the session. Thank you. Thank you for watching!

Agreement points

Need for inclusive and transparent AI governance frameworks grounded in human rights and constitutional values

Speakers

– Neema Iyer
– Mlindi Mashologu
– Matchiane Soueid Ahmed

Arguments

Need for AI that is based on accountability, trust, and a big component of public education

Need for transparent, inclusive AI governance frameworks grounded in constitutional values and public participation

AI governance must be co-created with African voices and grounded in human rights with civic oversight

Summary

All three speakers emphasize the critical importance of developing AI governance frameworks that are transparent, inclusive, and firmly rooted in human rights principles and constitutional values, with strong public participation and civic oversight.

Topics

Legal and regulatory | Human rights | Development

Importance of addressing bias in AI systems and ensuring diverse representation

Speakers

– Neema Iyer
– Mlindi Mashologu
– Lacina Kone

Arguments

Need for impact assessments, regional coalitions, investment in ethical and open-source alternatives that work within our context

Importance of removing bias from datasets and including diverse demographics in AI training

Need to preserve 2000+ African languages through locally trained AI systems to prevent cultural bias

Summary

Speakers agree on the critical need to address bias in AI systems through diverse representation in datasets, cultural preservation, and ethical alternatives that reflect African contexts and realities.

Topics

Human rights | Sociocultural | Legal and regulatory

Critical importance of data sovereignty and keeping African data under African control

Speakers

– Lacina Kone
– Matchiane Soueid Ahmed

Arguments

Over 1000 African startups daily downloading APIs from foreign models, with training data leaving the continent permanently

Data must remain in African hands as digital sovereignty is as important as territorial sovereignty

Summary

Both speakers strongly emphasize that data sovereignty is fundamental to African AI development, warning against the permanent loss of African data to foreign systems and asserting that digital sovereignty is as crucial as territorial sovereignty.

Topics

Legal and regulatory | Economic | Human rights

Need for blended financing approaches combining government, private sector, and international support

Speakers

– Lacina Kone
– Shikoh Gitau

Arguments

Private sector investment essential as mobile network operators with data centers have existential need to engage in AI

Need for blended financing approach combining government, private sector, and donor investments in compute infrastructure

Summary

Both speakers advocate for comprehensive financing models that leverage private sector capabilities, government support, and international partnerships to address Africa’s AI infrastructure and development needs.

Topics

Economic | Infrastructure | Development

Urgent need for AI education and digital literacy at grassroots level

Speakers

– Neema Iyer
– Lacina Kone
– Shikoh Gitau

Arguments

Critical need to democratize digital literacy and take AI conversations to grassroots level in accessible language

AI can reach indigenous people in rural areas and educate them in their own languages

Massive teacher training campaigns across six African countries showing high demand for AI education

Summary

All speakers agree on the fundamental importance of making AI education accessible to all levels of society, particularly in local languages and through creative, accessible approaches that reach rural and indigenous communities.

Topics

Development | Sociocultural | Capacity development

Similar viewpoints

Both speakers acknowledge the critical infrastructure gaps in computing power and digital connectivity across Africa as fundamental challenges that must be addressed for successful AI implementation.

Speakers

– Lacina Kone
– Mlindi Mashologu

Arguments

Africa lacks necessary computing power collectively across 50+ countries to train AI models locally

Digital divide and limited compute capabilities remain deep-rooted challenges requiring significant investment

Topics

Infrastructure | Development | Economic

Both speakers emphasize the importance of multi-stakeholder participation in AI governance, ensuring that diverse voices beyond government officials are included in policy development processes.

Speakers

– Neema Iyer
– Matchiane Soueid Ahmed

Arguments

Need for regional coalitions and bringing multiple stakeholders beyond just governments into AI policy discussions

AI governance must be co-created with African voices and grounded in human rights with civic oversight

Topics

Legal and regulatory | Human rights | Development

Both speakers advocate for proactive, intentional approaches to AI development that focus on practical utility and user needs rather than simply following global trends or being passive recipients of technology.

Speakers

– Lacina Kone
– Shikoh Gitau

Arguments

Africa needs deliberately intentional approach rather than passive optimism, focusing on user-centric AI models

Private sector must create enabling environments and massive awareness campaigns while governments provide supportive policy frameworks

Topics

Development | Economic | Legal and regulatory

Unexpected consensus

Strong agreement on the existential threat of data dependency on foreign AI systems

Speakers

– Lacina Kone
– Neema Iyer
– Matchiane Soueid Ahmed

Arguments

Over 1000 African startups daily downloading APIs from foreign models, with training data leaving the continent permanently

Critical need to track funding sources and understand foreign interests shaping Africa’s AI agenda

Data must remain in African hands as digital sovereignty is as important as territorial sovereignty

Explanation

Unexpectedly, speakers from different sectors (intergovernmental, civil society, and government) showed remarkable consensus on the urgency of addressing Africa’s dependency on foreign AI systems, with specific concern about the irreversible nature of data loss to external platforms.

Topics

Legal and regulatory | Economic | Human rights

Consensus on the need for practical, user-centric AI rather than pursuing the most advanced technology

Speakers

– Lacina Kone
– Shikoh Gitau
– Mlindi Mashologu

Arguments

AI should focus on agriculture, healthcare, and education as most useful applications for Africa’s development needs

Private sector must create enabling environments and massive awareness campaigns while governments provide supportive policy frameworks

Need for regulatory sandboxes and support for startups through AI-enabled economic transformation

Explanation

There was unexpected alignment across different stakeholder groups on prioritizing practical, contextually relevant AI applications over pursuing cutting-edge technology, suggesting a mature understanding of Africa’s development priorities.

Topics

Development | Economic | Sociocultural

Overall assessment

Summary

The speakers demonstrated remarkable consensus on key foundational issues including the need for inclusive governance frameworks, data sovereignty, addressing bias and cultural preservation, infrastructure investment, and education. There was strong agreement on the importance of multi-stakeholder participation and the urgency of addressing Africa’s dependency on foreign AI systems.

Consensus level

High level of consensus with significant implications for coordinated continental AI strategy. The alignment across government, civil society, private sector, and intergovernmental perspectives suggests a mature understanding of challenges and a shared vision for African-led AI development. This consensus provides a strong foundation for developing harmonized policies and collaborative approaches to AI governance across the continent.

Different viewpoints

Approach to AI policy development – harmonized continental framework vs. national sovereignty

Speakers

– Lacina Kone
– Audience from Gambia

Arguments

19 African countries have developed national AI strategies requiring harmonization while respecting sovereignty

Concern about countries working in silos rather than converging on common continental AI policy framework

Summary

While the audience member from Gambia advocated for a unified continental AI policy similar to the AU Cyber Security Convention, Lacina Kone defended the current approach of harmonizing diverse national strategies while respecting each country’s sovereignty, stating ‘not one size should fit all, but all sizes should fit together’

Topics

Legal and regulatory | Development | Economic

Primary responsibility for AI awareness and education

Speakers

– Shikoh Gitau
– Neema Iyer

Arguments

Private sector must create enabling environments and massive awareness campaigns while governments provide supportive policy frameworks

Critical need to democratize digital literacy and take AI conversations to grassroots level in accessible language

Summary

Shikoh emphasized private sector’s role in creating massive awareness campaigns and working directly with communities, while Neema focused on the need for democratized digital literacy through accessible education in local languages, suggesting different primary actors for education efforts

Topics

Development | Sociocultural | Capacity development

Unexpected differences

Limited explicit disagreement on fundamental AI risks and benefits

Speakers

– All speakers

Arguments

Various arguments about AI governance, infrastructure, and development approaches

Explanation

Surprisingly, there was minimal direct disagreement about the fundamental nature of AI risks or benefits. All speakers acknowledged both opportunities and challenges, with most disagreements focusing on implementation approaches rather than whether AI should be pursued or the severity of risks

Topics

AI Governance and Democratic Values in Africa | Development | Human rights

Overall assessment

Summary

The discussion showed remarkably high consensus on fundamental goals (inclusive AI governance, data sovereignty, capacity building) with disagreements primarily focused on implementation approaches, institutional responsibilities, and policy mechanisms rather than core objectives

Disagreement level

Low to moderate disagreement level. The speakers demonstrated strong alignment on overarching goals but differed on tactical approaches, suggesting a mature policy discussion where stakeholders share common vision but debate optimal pathways. This indicates positive potential for collaborative implementation despite methodological differences.

Partial agreements

Similar viewpoints

Both speakers acknowledge the critical infrastructure gaps in computing power and digital connectivity across Africa as fundamental challenges that must be addressed for successful AI implementation.

Speakers

– Lacina Kone
– Mlindi Mashologu

Arguments

Africa lacks necessary computing power collectively across 50+ countries to train AI models locally

Digital divide and limited compute capabilities remain deep-rooted challenges requiring significant investment

Topics

Infrastructure | Development | Economic

Both speakers emphasize the importance of multi-stakeholder participation in AI governance, ensuring that diverse voices beyond government officials are included in policy development processes.

Speakers

– Neema Iyer
– Matchiane Soueid Ahmed

Arguments

Need for regional coalitions and bringing multiple stakeholders beyond just governments into AI policy discussions

AI governance must be co-created with African voices and grounded in human rights with civic oversight

Topics

Legal and regulatory | Human rights | Development

Both speakers advocate for proactive, intentional approaches to AI development that focus on practical utility and user needs rather than simply following global trends or being passive recipients of technology.

Speakers

– Lacina Kone
– Shikoh Gitau

Arguments

Africa needs deliberately intentional approach rather than passive optimism, focusing on user-centric AI models

Private sector must create enabling environments and massive awareness campaigns while governments provide supportive policy frameworks

Topics

Development | Economic | Legal and regulatory

Key takeaways

Africa needs deliberately intentional rather than passive approaches to AI development, focusing on user-centric models that serve local needs in agriculture, healthcare, and education

AI governance frameworks must be co-created with African voices, grounded in human rights, and include transparent, inclusive processes with civic oversight

Data sovereignty is critical – African data must remain in African hands as digital sovereignty is as important as territorial sovereignty

Blended financing approaches combining government, private sector, and donor investments are essential for building necessary compute infrastructure

Multi-stakeholder collaboration is crucial, requiring involvement of civil society, youth, women’s rights organizations, trade unions, and other groups beyond just governments

Digital literacy and AI education must be democratized and delivered in local languages at grassroots levels to build public trust

Africa has over 2000 languages that need preservation through locally trained AI systems to prevent cultural bias and digital neocolonialism

Civil society plays essential roles as watchdogs, advocates, educators, and storytellers while documenting lived experiences of AI impacts

Resolutions and action items

Smart Africa to share benchmarking analysis of 19 national AI strategies with Council of ICT Ministers starting July

Smart Africa’s AI Council for Africa to focus on five areas: computing power, datasets, algorithms, AI governance, and market development

South Africa to leverage its G20 presidency to advance AI governance frameworks with focus on solidarity, sustainability, and equality

Continued teacher training campaigns across six African countries based on high demand demonstrated in pilot programs

Development of African AI Governance Toolkit (available at qbit.africa) and African AI Maturity Index (at datawall.africa) for government use

Implementation of mandatory public interest impact assessments before AI system deployment with ongoing monitoring

Establishment of regulatory sandboxes and startup support mechanisms for AI-enabled economic transformation

Unresolved issues

How to effectively harmonize 19 different national AI strategies while respecting individual country sovereignty

Addressing the fundamental infrastructure gap affecting 20% of population in remote areas across the continent

Preventing over 1000 African startups from continuing to use foreign AI models that permanently extract training data

Building sufficient collective computing power across 50+ African countries to train AI models locally

Operationalizing beautiful policies and frameworks – moving from high-level discussions to ground-level implementation

Addressing youth concerns about AI taking over jobs and building trust in AI technology among African populations

Determining realistic funding levels for Africa’s AI development compared to America’s $500 billion and France’s $200 billion commitments

Establishing clear definitions and metrics for what ‘democratizing AI’ actually means in African contexts

Suggested compromises

Blended financing model combining government, private sector, and donor investments rather than relying solely on public funding

Flexible policy harmonization approach where ‘each size should fit together’ rather than ‘one size fits all’ to respect sovereignty while enabling collaboration

Graduated approach to AI governance where countries can start with simple principles or policy frameworks rather than requiring full strategies immediately

Focus on ‘most useful’ rather than ‘most powerful’ AI applications suited to African development priorities

Parallel development of both enterprise-grade computing for private sector and accessible computing for researchers and startups

Government role focused on creating enabling environments and supportive policies while private sector leads innovation and awareness campaigns

We cannot be passively, we can no longer be passively optimistic. We have to be deliberately intentional… Africa today is the number one frontier in terms of availability of data to train AI… But they’re training those model. Those models are not located on our continent. They’re located outside the continent. And AI system is the way once you train the server, you can never get the information back again.

Speaker

Lacina Kone

Reason

This comment is deeply insightful because it reframes the entire AI discussion from a reactive to a proactive stance, while highlighting a critical paradox: Africa has abundant data but lacks control over how it’s processed. The irreversible nature of AI training he mentions introduces urgency to the sovereignty discussion.

Impact

This comment fundamentally shifted the conversation from theoretical policy discussions to concrete action items. It led other speakers to focus more on practical implementation and sovereignty issues, and established the framework for discussing the five pillars of the AI Council (computing power, data sets, algorithms, governance, and market).

Africa is not looking for the most powerful AI, it’s looking for the most useful one, looking at the agriculture, looking at the healthcare and looking at the education… AI allows us today to actually reach indigenous people in the rural area, to educate them in their own language, so they become tech-savvy like anyone who’s been to the university.

Speaker

Lacina Kone

Reason

This comment is thought-provoking because it challenges the dominant narrative of AI competition based on computational power and instead proposes a value-based approach centered on utility and inclusion. It redefines success metrics for African AI development.

Impact

This perspective influenced subsequent speakers to focus more on contextual applications and inclusive design. It helped establish a distinctly African approach to AI that prioritizes social impact over technological supremacy, which other panelists then built upon.

AI is already undermining governance in Africa, and we’re seeing this through automated disinformation campaigns, the eroding of public trust, manipulation of political discourse, and surveillance tools that are used to stifle voices at scale… we need AI that is based on accountability, trust, and a big component of public education.

Speaker

Neema Iyer

Reason

This comment is insightful because it grounds the discussion in current reality rather than future possibilities, providing concrete examples of AI’s negative impacts while maintaining a balanced view that acknowledges both benefits and harms.

Impact

This comment set a critical tone for the entire discussion, ensuring that subsequent speakers addressed both opportunities and risks. It influenced the conversation to consistently include safeguards, accountability measures, and the importance of civil society oversight in their responses.

Every time I hear democratizing this, democratizing that, I say, what’s the definition of democracy you’re talking about? And for me, democratizing means that it’s enabling everybody everywhere to have access to the same opportunities and resources… who is drafting these policies? What agenda do they have? Do they have Africa at heart when they are doing this?

Speaker

Shikoh Gitau

Reason

This comment is thought-provoking because it challenges the casual use of ‘democratization’ rhetoric and demands specificity about whose interests are being served. It introduces a critical lens about power dynamics in policy-making processes.

Impact

This comment deepened the analytical level of the discussion by questioning fundamental assumptions about who controls the AI narrative in Africa. It led to more nuanced discussions about agency, representation, and the need for African-led solutions rather than externally imposed frameworks.

We really need to democratize digital literacy. I feel like, you know, we’re in Oslo here having this conversation that doesn’t apply to most people living on the continent… We can’t always use these huge data governance languages, and then, you know, people’s eyes glaze over because they don’t really know what you’re saying.

Speaker

Neema Iyer

Reason

This meta-commentary is insightful because it critiques the very nature of high-level policy discussions while participating in one, highlighting the disconnect between elite discourse and grassroots reality. It calls for accessible communication and inclusive participation.

Impact

This self-reflective comment prompted other speakers to consider implementation and accessibility more seriously. It influenced the discussion to focus more on practical steps for community engagement and the need to translate policy into actionable, understandable terms for ordinary citizens.

Data must remain in African hands… sovereignty is not just territorial, but also digital.

Speaker

Matchiane Soueid Ahmed

Reason

This comment is thought-provoking because it expands the concept of sovereignty beyond traditional geographical boundaries to include digital assets, framing data control as a fundamental aspect of national independence in the AI era.

Impact

This comment reinforced and crystallized the sovereignty theme that ran throughout the discussion, providing a clear conceptual framework that other speakers could reference. It helped establish data sovereignty as a non-negotiable principle for African AI development.

Overall assessment

These key comments fundamentally shaped the discussion by establishing three critical frameworks: (1) the urgency of moving from passive to intentional action regarding AI governance, (2) the need to define African success metrics based on utility rather than power, and (3) the importance of questioning who controls the AI narrative and ensuring African agency. The comments created a progression from identifying current harms to proposing African-centered solutions, while consistently challenging assumptions about democratization, sovereignty, and inclusion. The discussion evolved from theoretical policy considerations to practical implementation strategies, with each insightful comment building upon previous ones to create a comprehensive vision for African AI governance that prioritizes local needs, democratic values, and genuine self-determination.

How do we operationalize beautiful policies and frameworks into actual ground-level implementation?

Speaker

Neema Iyer

Explanation

This addresses the critical gap between policy development and practical implementation, which is essential for ensuring AI governance frameworks actually work in practice rather than just existing on paper.

How can we develop funding models for AI innovation that are sustainable and bring young people into product creation?

Speaker

Neema Iyer

Explanation

This highlights the need for research into alternative funding mechanisms that support local innovation and youth engagement, moving beyond traditional donor-dependent models.

What are the foreign interests and agendas of tech giants in shaping Africa’s AI agenda?

Speaker

Neema Iyer

Explanation

Understanding the motivations and influences of external actors is crucial for developing truly sovereign AI strategies that serve African interests rather than external ones.

How do we move AI governance from paper to pavement – from policy documents to actual implementation?

Speaker

Shikoh Gitau

Explanation

This emphasizes the implementation gap that exists between developing AI policies and actually deploying them effectively in real-world contexts.

What realistic steps can governments take to ensure AI readiness, particularly regarding digital literacy and infrastructure?

Speaker

Peter King (Audience member from Liberia)

Explanation

This calls for concrete, actionable research on practical steps governments can take to build foundational capabilities needed for AI deployment.

How can we create convergence across African countries for AI policy instead of working in silos?

Speaker

Alhaji (Audience member from Gambia)

Explanation

This addresses the need for research into harmonization mechanisms that respect national sovereignty while enabling continental coordination.

What specific measures can build trust among African youth regarding AI and job displacement concerns?

Speaker

Lydia Lamisa Akamvareba (Audience member from Ghana)

Explanation

This highlights the need for research into youth perceptions of AI and effective strategies for building confidence rather than fear about AI’s impact on employment.

How do we assess and measure AI maturity across different African contexts?

Speaker

Shikoh Gitau

Explanation

Understanding where countries stand in terms of AI readiness is essential for developing targeted interventions and tracking progress over time.

How can we ensure that AI models respond to African healthcare needs and realities?

Speaker

Shikoh Gitau

Explanation

This points to the need for research into developing and benchmarking AI systems that are specifically designed for African contexts and needs.

How do we completely change assessment methods in education given AI’s impact on traditional evaluation?

Speaker

Neema Iyer

Explanation

This highlights the need for research into new educational assessment frameworks that account for AI tools and changing skill requirements.