African Countries Are Racing to Create AI Strategies — But Are They Putting the Cart Before the Horse?

African countries are increasingly establishing policy frameworks for emerging technologies such as Artificial Intelligence (AI). In the last seven years, at least eight African countries have adopted national AI strategies, while five have either completed a draft or are in the process of developing one (Figure 1). Our recently published report with the Atlantic Council’s Digital Forensic Research Lab (DFRLab) discusses in detail the key trends in emerging tech governance on the continent, focusing on AI, digital public infrastructure (DPI) and connectivity in five countries — South Africa, Kenya, Nigeria, Ghana, and Zambia.

The report highlights that emerging tech governance in Africa feeds into broader continent-wide sentiment of harnessing technology for development. This approach is further accentuated by the recently published African Union (AU) Continental AI Strategy, which highlights AI’s role in driving the AU Agenda 2063 and the UN SDGs. Discussions about digital infrastructure expansion, talent development, public sector capacity building, digital literacy, and private sector investment drive home this optimism. 

This trend continues to rise. In the last six months, four countries — Zambia, Egypt (2nd edition), Lesotho, and most recently, Kenya — have published AI strategies. There are important justifications for having stand-alone AI strategies: it signals government commitment to the sector and helps outline clear priorities for resource allocation toward AI infrastructure, talent development, and capacity building. Explicit AI strategies also provide an opportunity to set clear ethics and regulation guidelines for AI development, creating certainty for investors and demonstrating that countries have a stake in global AI governance discussions. Given the increasing centrality of AI technologies to the global economy, countries looking to boost competitiveness might also use AI strategies to articulate an industrial policy for the industry.  

Figure 1: African countries with AI strategies.

Source: Authors

But to what extent do these AI strategies reflect respective African countries’ digital economy, and effectively address persistent challenges such as human capital deficits, infrastructure limitations and financial constraints? To understand this, this article provides a brief summary of AI strategies from two of the countries covered in our DFRLab Report: Kenya and Zambia, and then discusses some of the characteristic shortcomings of African national AI strategies. 

Kenya’s National AI Strategy 

Launched in March 2025, Kenya’s National AI Strategy takes a comprehensive outlook toward strengthening Kenya’s AI ecosystem, emphasising the economic, social, and political dimensions of its development. Similar to the AU Continental AI Strategy, it prioritises AI adoption in critical sectors such as healthcare, agriculture, education, security, public service delivery, and SMEs, and outlines a phased approach towards implementing its core objectives.  It also has a strong focus on issues of ethics and human rights — as defined by its guiding principles of inclusivity, co-creation, transparency and accountability, local-first, environmental sustainability, among others. 

The overarching ambition of the strategy is to position Kenya as a regional leader in AI research and innovation by leveraging its existing strength in other digital innovation domains. Kenya is already one of the top five African tech hub nations and is a leading destination for VC investments in AI on the continent. Growing foreign investments in AI relevant infrastructure — for example, the recent partnership between the USTDA and Kenyan Semiconductor Technologies Limited to establish a semi-conductor manufacturing facility in Nairobi — and the increasing presence of big tech AI research labs strongly position Kenya to achieve this goal. 

The AI strategy outlines seven outcomes for achieving this ultimate goal. These include modernising Kenya’s digital infrastructure, building an innovative and sustainable data ecosystem, upscaling R&D for developing advanced  indigenous AI models, upskilling the AI talent base, creating an agile and adaptable regulatory environment, expanding public and private investment in AI, and strengthening ethics and inclusivity practices in AI development. 

Figure 2: Kenya's AI strategy house

Source: Kenya's National AI Strategy , 2025-2030

A few things are unique about the Kenyan AI Strategy. First, it builds on a baseline evaluation of Kenya’s current AI landscape, drawing on the UNESCO Readiness Assessment (RAM) earlier conducted in the country, and providing detailed recommendations based on appraisal of Kenya's comparative advantages. The Strategy also uniquely anticipates potential challenges viz a viz the strategic opportunities available in Kenya’s AI industry. A SLOC (Strength, Limitations, Opportunities and Challenges) analysis assesses the potential impacts of labour force disruptions; digital divide; data sovereignty; ethics, trust, and human rights issues; weak regulatory landscape; equitable access to AI-driven public services; and environmental sustainability of AI, and provides strategic recommendations to address these. Lastly, its development process follows a comprehensive stakeholder engagement that includes key informant interviews, focus group discussions, townhall meetings, and online surveys to gather perspectives from a broad spectrum of experts. 

While Kenya boasts of a budding digital ecosystem, current metrics show weak public sector leadership in responsible AI, evidenced in a particularly low performance on the Government Actions (4.68) indicator of the 2024 Global Index on Responsible AI. Establishing an enabling regulatory environment is crucial in driving society-wide enthusiasm toward AI development and encouraging responsible usage and adoption. 

Zambia’s National AI Strategy 

Zambia’s National Artificial Intelligence Strategy (2024 - 2026) was officially launched in November 2024 in a process led by the Ministry of Technology and Science through consultations with diverse stakeholders, technical assistance from the Tony Blair Institute, and contributions from the Government of Finland and USAID Open Spaces . The implementation is planned to be executed in three phases: the first 100 days, the first year, and the second year, each with specific objectives and activities. 

The strategy aims to integrate AI into key priority domains: healthcare, agriculture, education, mining, climate, and public services, leveraging AI technologies to drive sustainable and inclusive development. It envisions making Zambia a “destination for ‘AI for Emerging Economies’ ventures”, fostering an environment that would enable  AI-driven innovations that address low- and middle-income countries’ challenges. 

Zambia’s AI Strategy is committed to global best practice frameworks including the AU continental AI Strategy, UNESCO Recommendation on the Ethics of Artificial Intelligence and the OECD Principles on Artificial Intelligence. A comprehensive analysis performed evaluated existing infrastructure, skills, and overall digital environment to identify the key challenges and opportunities as illustrated below: 

Figure 3: Challenges and corresponding strategic objectives of Zambia's National AI strategy

Source: Authors, based on information in Zambia's National AI Strategy

The strategy envisions realising these objectives through a robust governance framework, bridging persisting gaps in digital infrastructure, promoting human capital and investment, and harnessing international collaborative research and innovation networks for fair and ethical use of AI. 

It proposes the establishment of a National AI Council, an expert group to offer high-level advisory, strategic guidance and stakeholder engagement on AI-related matters for government decision making. Sector-specific Technical Working Groups will also be established to tackle technical challenges and oversee implementation within key sectors. A National Emerging Technologies Centre of Excellence is proposed to facilitate international cooperation on AI to mobilise technical and resource support, facilitate knowledge exchange, and ensure adherence to global standards. 

Existing institutions are also integrated into the plan: the Ministry of Technology and Science will be responsible for integrating AI across the economy, and the Smart Zambia Institute is now responsible for integrating AI into public administration and service delivery. The strategy also extends the regulatory mandate of the Zambia Information and Communications Technology Authority to include AI regulation and ethical compliance. 

Zambia ranked low (92 out of 138 jurisdictions) in the Global Index on Responsible AI (2024), and scored particularly poorly in the Responsible AI Governance dimension (1.93) and the Governance Framework pillar (2.60). Therefore, establishing a governance body, along with a centre paying particular attention to AI ethics  are positive steps toward responsible AI governance in Zambia.

What are the cracks in the road? 

Before discussing some crucial issues left unaddressed in these strategies, it is important to acknowledge the unique contributions they both offer. For AI strategies to be effective, specific measurable goals in the form of detailed key performance indicators (KPIs) are important. The Kenyan AI Strategy demonstrates this in some ways. For example, it provides time-bound recommendations for developing an indigenous AI ecosystem, such as establishing AI hubs within special economic zones (SEZs). This contrasts with current recommendations, which often suggest regurgitated policy options such as “establishing an AI fund” or generic solutions like “public-private partnerships”, which lack clear implementation blueprints. 

In both the Kenyan and Zambian strategies, the assessments conducted ahead of developing the policy document were essential to understanding the countries’ key bottlenecks and designing a strategy that responds to their respective realities. Both also involved extensive stakeholder engagement, which is crucial for driving policy uptake at the local level and encouraging ownership from a wide spectrum of the AI community. Beyond national frameworks and coordination efforts, global networks can be leveraged to address financial and capacity limitations through funding and investment frameworks as well as knowledge and resource sharing schemes. Both strategies have made a strong start in this regard by engaging international partners’ expertise in the policymaking process, and referencing established ethical guidelines to ensure its compatibility for global partnerships. 

However, similar to other African countries’ national AI plans, these two strategies fail to address several issues. We discuss a few below:

The race against unrealistic timelines: One key issue with Zambia’s strategy in particular is that the strategy’s goals are too ambitious to achieve within a relatively short timeline. Addressing foundational issues such as reliable power systems and connectivity, and bridging the urban-rural digital divide that ensures inclusive development is built-in instead of an afterthought will already require a significant amount of time. 

High-stakes sectors require stronger regulation: The prioritisation of key domains such as healthcare and education presents higher risks that could potentially exacerbate inequalities and expose vulnerable populations to unprecedented dangers necessitates a strong regulatory environment to ensure policy enforceability. In particular, the question of how governments will ensure compliance in a way that they have not been able to for the preceding Data Protection Act comes to mind. 

Lack of specificity in benefit-sharing mechanisms: Another limitation is the  lack of specificity and concreteness regarding benefit-sharing mechanisms both within the countries, and in international partnerships. Despite emphasising inclusivity, most African national AI strategies fail to provide details on specific, committed mechanisms (beyond general goals or pilot considerations) to ensure broad socio-economic benefit-sharing from AI-driven productivity gains, which is vital for social stability. 

Undefined roles in Global Governance: Cooperation and participation in global AI governance are often emphasised in African countries’ national AI strategies. But these do not provide clarity on their intended  responsibilities within global AI governance. For example, the Kenyan strategy mentions international engagement in positioning Kenya as a regional AI leader, but fails to define a more specific role for Kenya in global collaborations and policy advocacy focused distinctly on the technical safety research and governance norms for advanced AI systems. More ambitiously—though acknowledging constraints—the strategy would have fronted ideas or called for an African AI safety and security institute as a “rallying call” to other African countries.

Confronting structural challenges: More broadly, it is important to recognise that Africa’s development challenges are not only a result of slow technological advancements, but also deeply rooted systemic issues including exploitative resource extraction, asymmetric geopolitical power structures and epistemic injustices, which may undermine AI technologies’ transformative potential. Tackling the key challenges highlighted across African countries’ AI strategies, particularly infrastructure and talent development, will not only require monetary investment, but also extended time and continuous effort from various actors. The pace at which such transformational change can occur may not keep up with how fast AI evolves. Although there are plans for attracting funding, it is unclear how countries would be able to attract capital, and whether the dire need for investment may come at the expense of ethical considerations in developing and deploying AI tools in Africa.  Moreso, international actors' involvement in strategy development and skills training programs raises concerns beyond monetary dependence, such as whose priorities are guiding Africa’s future in AI and whether they are representative of Africans’ desires and needs. 

Rethinking capacity building: There is also a strong focus on capacity building and retaining local talent focused on technical AI skills — and rightly so — but cultivating skills such as deep ethics, creativity, critical oversight should also be prioritised, as they are crucial for ensuring long-term human safety and societal wellbeing. 

Ecological limits and AI’s environmental costs: Sustainability is also merely mentioned as a cross-cutting issue, however, they do not address the potential natural resource and energy demands of high-performing AI. Acknowledging ecological limits as hard constraints is crucial for long-term safety and security.

The limits of multi stakeholder engagement: While it is commendable that some strategies have employed multistakeholder engagement, expert discussions have noted that only few elite groups are informed about the impacts of emerging technologies, and engaged in consultative platforms, which risks exacerbating digital inequities within the continent. 

Premature focus on AI?

Many African countries are building AI strategies on top of existing digitalisation ambitions that have remained unrealised. In particular, the overemphasis on “integrating AI into various sectors of the economy” crowds out the role that other unique digital technologies could play or are currently playing in socioeconomic development. 

The world today is embracing an AI buzz, but its attention in the future might turn to another technology domain. African countries must ensure they are not prematurely focusing on AI strategies, while many other foundational considerations are left unaddressed. Instead, using relevant regulatory instruments to address AI-specific issues such as data governance, liabilities and risks, while developing broader technological advancement strategies in alignment with national development plans will ensure policy stability, provide informed foresight into how new technologies can be integrated into the larger goal of addressing socio-economic challenges, and promote collective efforts for ethical and responsible strategies for technology-enabled development. 

We are grateful to George Gor for comments on the draft of this piece. 

Authors: Selam Abdella and Ayantola Alayande

Acknowledgement:

This analysis is based on research funded by the International Development Research Centre (IDRC) and UK International Development under the AI4D program, as part of the African Observatory on Responsible AI.