Artificial intelligence (AI) is shaping the lives of individuals, economies, and countries worldwide. Innovations associated with AI are advancing across sectors such as health, transportation, and agriculture, transforming economies and entire business models and practices. In Africa, AI has the potential to grow the continent’s economy by an estimated $2.9 to 4.8 billion by 2030. In recognition of the AI promise, African stakeholders are increasingly positioning themselves to expedite AI adoption and realize its benefits. This is most evident through their pursuit of AI innovations, infrastructure development, governance, and convenings like the prominent inaugural Global AI Summit on Africa in Rwanda on April 3–4, 2025.
This summit in Kigali was hailed as a momentous occasion for the continent, as it brought together policy leaders, private sector representatives, and other stakeholders from across the continent to forge a collective pathway on shaping Africa’s role in the global AI economy. The gathering resulted in the Africa Declaration on Artificial Intelligence, which received endorsements from forty-nine African countries, the African Union, and Smart Africa. The declaration affirms the continent’s vision on AI, as outlined in the African Union (AU) Continental AI Strategy released in 2024. It makes commitments to grow seven key areas: talent, data, infrastructure, market, investment, governance, and institutional cooperation. The declaration also announced the creation of a $60 billion Africa AI Fund and an Africa AI Council. The council will be pivotal in promoting AI initiatives throughout the continent, particularly in governance.
Globally, countries are engaging in AI governance due to both the rapid growth and transformative potential of the technology as well as concerns about societal implications. As AI systems become more pervasive, leaders have called for greater international cooperation to govern the systems’ development and deployment. Examples of this cooperation include the United Nations High-Level Advisory Body on AI, which developed the “Governing AI for Humanity” report; a series of AI safety summits; and the International AI Safety Report. However, there is no universal model for AI governance. A World Bank report details four current approaches: industry self-governance, soft law (such as strategies, policies and standards), hard law (such as the European Union AI Act), and regulatory sandboxes.
In examining Africa’s evolving landscape of AI governance, there is significant evidence of AI soft law, particularly through the development of national AI strategies and policies. This policy outlook highlights the continent’s progress through exploring two perspectives: policy practice and policy discourse. The analysis of AI-specific policy documents and a high-level dialogue on the sidelines of the Kigali summit indicate that tangible progress is being made through strategies and policies that promote the development of localized AI applications to provide socioeconomic benefits.
AI Policy Practice: What Data Show
Research conducted by the Africa Program at the Carnegie Endowment for International Peace, through the Africa Technology Policy Tracker (AfTech) tool, reveals that African countries are at different phases of AI governance, and this process is evolving through the formulation of national AI strategies or policies. There are fifteen national and two continental AI strategies and policies published to date (see Figure 1).
Among the national documents published, there is a distinction between AI strategies and AI policies. Currently, there are twelve national AI strategies and three national AI policies. While all these documents show country direction for AI development, there is a notable preference for AI strategies. An AI strategy typically aims to establish direction and answer the “how” by providing a road map to harness AI for socioeconomic gain, whereas a policy usually establishes more specific guidelines and principles for technological developments. However, this distinction is not always clear, as the analysis of Africa’s national AI strategies and policies shows comparable approaches, with both types of documents providing vision and guidance on how to effectively harness AI for socioeconomic benefit.
The continental policy frameworks include the AU Continental AI Strategy and Smart Africa’s AI Blueprint. The AU strategy is regarded as the anchor AI document, as it provides a singular vision for the union’s fifty-five member countries. It identifies fifteen recommendations including the development of national AI strategies and policies, calling them “important starting point[s] for governing AI.” Notably, there are other digital policies and legal frameworks used to govern AI. For instance, cyber security laws, cloud policies, and data protection regulatory frameworks are complementary AI governance documents. But AI-specific policy documents are generally considered the most effective for coordinating the development, deployment, and governance of AI.
The year 2024 was a pivotal year for AI governance on the continent. Six AI-specific documents were published: five at the national level by Ethiopia, Libya, Mauritania, Nigeria, and Zambia and one at the continental level (the Continental AI Strategy), together with white papers. The year 2025 is looking like a year for enhanced AI policymaking as well. In the first quarter alone, Côte d’Ivoire, Kenya, and Namibia published national strategies, while countries such as Lesotho and Tanzania have released draft strategies. Collectively, these documents provide key insights into how Africa is positioning itself in the AI age. They reveal two overarching priorities: leveraging AI to advance digital transformation and investing in the rapid development of domestic AI capabilities.
AI for digital transformation: African AI strategies and policies have been developed to help drive the continent’s digital transformation. For example, the AU Continental AI Strategy positions AI as an enabler of socioeconomic growth and aligns it with the aspirations of the AU’s Agenda 2063, which are achieving both inclusive and sustainable socioeconomic development for all Africans. The strategy emphasizes the significant potential of AI to transform the health, education, and agriculture sectors. This framing is also reflected in national documents. African countries have designed national strategies and policies that identify priority areas for AI application to provide socioeconomic benefits. While analysis using AfTech reveals several common priority sectors of AI application—agriculture, government, and health services (see Figure 2)—each country has its own unique list of priority sectors, which are identified as having high potential and impactful use cases. Various criteria are used to select sectors, which typically align with the national development agenda. For example, the priority use cases in Kenya’s AI Strategy align with the government’s Bottom-Up Economic Transformation Agenda and its Vision 2030. Egypt identifies AI use cases based on three primary factors: feasibility, which pertains to technology maturity; desirability, focusing on the impact on society; and viability, concerning the return on investment. Similarly, Rwanda’s selection was made by identifying sectors where AI can produce significant outcomes and benefits.
Building AI enablers: Alongside identifying potential sectors for AI adoption, African policy practice identifies the necessary enablers or capabilities required to develop AI applications. AI enablers, often referred to as capabilities or pillars, represent the essential inputs or foundational infrastructure required to foster and facilitate AI development and adoption. Analysis indicates that governance and talent are fundamental enablers common across all African countries, followed by data, research and innovation, and digital infrastructure (see Figure 3). Identifying the enablers is critical, as it helps guide the policy action through goal setting and road maps to develop AI capabilities. Financing and the vehicles for actualization—either through government investments or through public-private partnerships—are also important enablers.
AI Policy Dialogue: Insights From Policymakers
To further explore and understand Africa’s evolving policy landscape, Carnegie’s Africa Program, in partnership with Smart Africa, organized a ministerial dialogue at the Global AI Summit on Africa. The dialogue yielded significant insights that supported the above findings on policy practice and introduced fresh perspectives on Africa’s AI policy practice.
AI for Africa by Africa
African policymakers aim to ensure that AI development is aligned with the continent’s needs and accurately reflects the perspectives of its people. During the dialogue, stakeholders noted that much of Africa’s AI is shaped by external influences, and there is an imperative to change this narrative. A key concern is that AI technologies in Africa are predominantly imported, which means that they may not be designed to address the unique needs and contexts of the continent. Furthermore, many global conversations and outcomes on AI have been shaped without consideration of regions like Africa. Upending this narrative was highlighted as an urgent imperative that needs to be addressed through policy prescription.
One speaker remarked, “AI should not be built for Africa—but with Africa, by Africa, and for Africa.” Another speaker emphatically stated, “It must be in the hands of Africans to determine how AI is built and governed to serve Africa.” And yet another speaker stated, “Africans are done with letting the world dictate the terms of an AI race in which they are meant to ‘catch up.’”
A distinct message emerged: Africa must take charge of its AI journey, and this should be done through AI policy frameworks that provide clear and actionable road maps. One participant remarked:
Africans are ready to redefine this race for themselves, including what the finish line should and should not be. Maybe this means aiming for local models that outperform imported models in how well they respond to and serve African populations, including across the rich diversity of languages spoken throughout the continent. Maybe it means building AI that can do more with less in a more sustainable way, consuming less energy while demanding less cutting-edge compute. Maybe it means investing in the next wave of global AI talent, given that by 2050, one in four people in the world will be African. Or maybe it means arriving at a continentally harmonized enabling environment that allows African AI to thrive in a single digital market.
AI for Good
Another insight from the dialogue is that AI governance in Africa is designed to provide socioeconomic benefits to Africans. Participants emphasized that AI should be viewed as a means for digital transformation to address socioeconomic inequalities and unemployment challenges, particularly among the youth. Participants observed that Africa’s population is predominantly youth, but there is a scarcity of jobs and economic opportunities. Africa’s AI moment must offer its youth an opportunity to rewrite their story, with tangible digital dividends. This should be accomplished through policy interventions that harness the creative abilities of young people and foster environments where innovation can flourish.
Developing AI Enablers
The dialogue also highlighted the importance of advancing talent, data, governance, research and innovation, and digital infrastructure, all identified as essential AI enablers. Participants observed that there is a shortfall in all five , which hinders the continent’s capacity to develop localized AI. The following points were raised regarding their establishment across Africa.
Talent
Participants emphasized the importance of investing in AI talent to enhance both basic literacy and advanced skills. It was noted that only 3 percent of the global AI talent pool resides in Africa, even though the continent boasts one of the largest and fastest-growing young populations. The youth bulge presents the continent with an incredible opportunity to leverage AI, but it is essential to ensure that this demographic is adequately skilled. Government representatives emphasized current initiatives, notably one in Togo, which seeks to train 50,000 individuals in AI skills by 2025, with intentions to train an additional 50,000 individuals each subsequent year. Participants also highlighted the necessity of investing in the broader AI ecosystem alongside AI skills, as insufficient demand for these skills will contribute to brain drain.
Data
Data serves as the foundation for AI innovation and development, and a key concern is that current large language models are not built on African data and that much of the continent’s data are owned by foreign entities—especially big tech firms—raising critical issues of data sovereignty. Participants emphasized that the creation of localized data sets that incorporate African languages is a priority. To develop “AI for Africa” that reflects African values, contexts, and realities, local datasets are essential. It was also noted that the absence of African data has greatly contributed to AI bias. In recognition of this issue, a participant explained how the Nigerian government is collaborating with academia and the private sector to build a multilingual large language model (LLM) in five local languages to drive AI inclusion. Noteworthy is that multiple AI strategies—such as those of Ethiopia and Senegal—prioritize developing AI that incorporates and promotes local languages. Participants also highlighted the necessity of establishing cross-border data transfer frameworks to foster a supportive, AI data–driven regional environment and called for member states to adopt the AU Data Policy Framework to help address issues of data localization, classification, and cross-border transfers.
Governance
Participants observed that AI governance in Africa is in its early stages; however, the formulation of strategies and policies represents a deliberate attempt to leverage the advantages of the technology while addressing potential risks. Four issues were identified in AI governance: regional harmonization, safety, sovereign AI, and an innovation enabling environment.
Regarding harmonization, it is essential for countries to engage in intentional efforts to align their AI policy practice with the AU Continental AI strategy. As a participant noted, “Each African nation must chart its own AI destiny while contributing to a collective vision.” Harmonization efforts are crucial for avoiding policy fragmentation and duplication and for supporting the African Continental Free Trade Area, as well as providing the continent with a unified and influential global voice. Participants, however, cautioned against “copy-pasting,” emphasizing that policy frameworks should be customized to align with national agendas and realities.
On safety, participants stressed that governance should facilitate safe, inclusive, and responsible advancement of AI. This should be done through a multisector approach and through subsidiary legislation. Furthermore, discussions advocated the establishment of regional standards by the AU to enhance responsible and transparent AI practices.
Regarding AI sovereignty, one participant stated, while indicating a sense of urgency, “We can’t afford to treat AI like any other technology. It’s a technology that increasingly shapes how people live and how people think, and so its creation and governance must be recast as a matter of sovereignty.” Sovereign AI involves the development of local AI models along with the necessary infrastructure, including local computing and data resources. Currently, most of Africa’s AI is imported, meaning that it is modeled though external actors. In a world where technology is increasingly becoming a geopolitical point of contention, African policymakers are increasingly concerned about undue influences and Africa’s agency in AI development.
Lastly, it was observed that AI governance should promote an enabling environment that balances overprotection and innovation.
Research and Innovation
Participants noted that there is limited AI research on the continent, which impedes AI innovation and the creation of use cases tailored to the African context. Policy interventions are needed to increase the number of researchers and developers focusing on continental issues. Furthermore, African countries must cultivate entrepreneurship ecosystems that encompass research universities, startups, venture capitalists, and access to markets and opportunities on both regional and global scales. In summary, the advancement of AI necessitates a holistic approach and cooperation among diverse stakeholders.
Digital Infrastructure
The advancement of AI depends on robust computing resources, consistent electricity supply, high-speed internet, and adequate data storage options—all of which African countries are currently lagging behind on. Across the continent, the shortage of graphics processing units is particularly debilitating, as they are essential for training large AI models. Currently, Morocco and South Africa are home to the most advanced supercomputers capable of undertaking complex computational tasks needed for AI development. Relatedly, data centers are also lacking across the continent. In 2024, data centers in Africa represented less than 1 percent of the global total. A recent study shows that most powerful data centers are in China and the United States, indicating an AI divide that needs to be mitigated. The lack of AI infrastructure in Africa means that its countries have limited capacity to build and train advanced AI models, an impediment to sovereign AI ambitions. Investing in these resources is paramount, but, of course, AI infrastructure requires large financial investments as well as complementary resources such as energy.
Financing AI: Collaboration and Partnerships
Participants called for holistic and innovative strategies, such as whole-of-government approaches, pooled resources, regional investments, public-private partnerships, collaboration with the African Diaspora, and partnerships between the private sector, particularly between Western companies with advanced AI capabilities and African tech firms. An example of the latter emerged during the Kigali summit: the Africa-based global tech company, Cassava Technologies, will invest up to $720 million to work with the U.S.-based global tech company, Nvidia, on building the first “Africa AI Factory.” The investment will provide AI infrastructure to Egypt, Kenya, Morocco, Nigeria, and South Africa.
The significance of pooled resources was highlighted as a means to address the scale and diversity of AI resources required on the continent. Estimates indicate that more than $7 billion is required to adequately address deficiencies in data, computing, and skills across the continent. While a $60 billion Africa AI Fund was announced as an outcome the summit, demonstrating a strong commitment to pooling resources, details of the fund are yet to emerge.
Collaboration with the African diaspora was also underlined as a significant resource for advancing AI in African nations, particularly regarding talent and knowledge sharing. Nigeria, for example, highlighted how its diaspora community served as a vital knowledge partner in shaping its AI strategy. Policymakers noted that as countries pursue collaborations and partnerships, they should exercise caution in order to maintain agency (not be boxed in) and safeguard digital sovereignty.
Conclusion
Africa stands at a pivotal juncture in shaping its AI future—from being largely a consumer of technologies to being an active architect of its own digital destiny. Significant gains have been made in its policy posture but more needs to be done, especially in policy action. As the excitement of the inaugural Africa AI Summit fades, it is critical that the momentum gained continues within the continent and beyond. To achieve the vision of “AI for Africa by Africa,” African governments and stakeholders must move decisively—from strategy to implementation, from vision to innovation, and from dialogue to measurable impact. The insights from both policy practice and dialogue underscore a clear imperative: AI governance in Africa is bold and purpose driven. It prioritizes building national AI capabilities, growing AI applications, and fostering inclusion and sovereignty, and seeks to translate vision into action through strategic policy action. As the continent moves from declarations to delivery, the real measure of success will be whether AI can be developed with real use cases that serve Africans (particularly youth), enhance lives, and move Africa from consumer to global leader in the AI age.
