Speaker 1 : Dr Aristide Akem, University of Southampton, Southampton, United Kingdom
Title: Can Africa build efficient AI using edge/cloud federation?

Abstract: Africa has the opportunity to rethink how AI infrastructure is built. Rather than trying to reproduce hyperscale models developed elsewhere, distributed architectures that combine edge devices, local infrastructure, regional clouds, and selective access to foundation models could better address the continent’s realities. Through examples of emerging AI initiatives for African languages and local services, this talk discusses how such approaches can improve efficiency, affordability, and digital sovereignty, while opening new pathways for innovation tailored to the African context.

Speaker 2: Prof. Mouhamadou Lamine BA, Ecole Supérieure Polytechnique, Université Cheikh Anta Diop, Dakar, Senegal
Title: From Consumer to Innovator: Towards African Technological Autonomy in Artificial Intelligence
Africa as an AI consumer or Africa as an AI innovation laboratory.

Abstract: Africa currently finds itself in a position of passive consumer of AI solutions developed by other continents. This technological dependency poses structural risks: imported solutions, designed for external contexts, often fail to address the specific and nuanced challenges of the continent. They perpetuate a form of technological colonialism that limits African agency and concentrates value creation elsewhere.
The figures of dependency: The AI market in Africa is estimated at USD 4.51 billion in 2025, with a projected growth to USD 16.53 billion by 2030 (CAGR of 27.42%). Yet this growth primarily benefits foreign technology companies. Data generated on the continent—estimated at several billion data points annually—is extracted by multinationals without significant returns for Africa. Moreover, Africa represents only 2-3% of global AI research, despite its unique needs.
This presentation argues that Africa must execute a strategic transition: shifting from the role of consumer to that of AI innovation laboratory. By drawing on lessons learned from global experiences, the continent possesses the potential to develop frugal, contextualized, and resilient AI solutions adapted to local realities—limited infrastructure, heterogeneous data, unique sectoral needs (agriculture, healthcare, inclusive finance).
The emerging ecosystem: Africa currently hosts 207 AI startups (99% growth since 2022), with concentration in South Africa, Nigeria, and Kenya, but growing diversification toward Egypt, Tunisia, and Ghana. Innovation laboratories such as the AI4D Network of Labs (operating in 11 African countries), the Google Africa Applied AI Lab (based in Accra), and the Centre for Artificial Intelligence Research (South Africa) are developing contextualized solutions in healthcare, agriculture, language technology, and education. The AI Hub for Sustainable Development (launched in 2025) supports 130 African startups with access to high-performance computing infrastructure.
We will explore the nature and implications of this current dependency, map the emerging landscape of AI innovations already present on the continent, and identify strategic levers to accelerate this transition toward African technological autonomy.

Speaker 3 : Prof Jean Louis FENDJI, University of Ngaoundere, Ngaoundere, Cameroon
Title: How Could AI Work with African Data?

Abstract: Only about 2% of the world’s AI training data comes from Africa. The standard fix – more hackathons, more annotation drives, more collection campaigns – cannot close that gap. You cannot annotate your way to a trillion tokens.
The missing piece is not datasets; it is datafication: the continuous data streams that digital infrastructure generates as a by-product of ordinary life. Tellingly, not one of nine African AI strategies analysed even uses the word. Policies cannot invest in what they cannot name.
Africa should not copy the Global North’s big-data machinery either – too costly, too power-hungry, too wasteful. The opportunity is better data, not bigger: a health worker’s phone turning every clinic visit into a record; a $50 solar sensor streaming three seasons of farm data; communities owning those streams.