ARUA-CD event examines the role of AI in advancing climate research in Africa

At this year’s ARUA Biennial Conference hosted by Makerere University, the ARUA Centre of Excellence in Climate and Development (ARUA-CD) convened a dedicated session exploring how artificial intelligence (AI) and machine learning are shaping climate research across Africa. The session brought together early career researchers, senior academics and practitioners to discuss practical applications, emerging challenges and the wider implications for research capacity on the continent.

The event formed part of the conference’s broader emphasis on research and innovation as drivers of Africa’s development, particularly as higher education institutions seek new ways to address sustainability, resilience and technological transformation.

Navigating Interest, Enthusiasm and Key Skill Gaps

The session opened with a real-time Mentimeter exercise inviting participants to share what they want to understand most about AI in climate research. Themes such as access to relevant datasets, developing technical skills, navigating ethical issues and translating AI outputs into policy emerged strongly. 

Part of the challenges faced by early careers on the African continent in adopting AI tools as discussed was the challenge of scarcity of data to enable the AI tools make the right predictions. This also makes climate change modelling a great challenge. Another challenge discussed was how early careers can get mentoring in AI research,” explained Ezekiel Acquaah, research fellow at the University of Ghana.

Presentations Illustrate the Range of AI Applications

A series of five short presentations provided insights into how AI is already being used across different areas of climate research and governance.

• Accelerating evidence synthesis - From the University of Cape Town, Nicholas Simpson presented examples of machine learning tools used in large evidence syntheses. He highlighted how automation can help researchers process [U1] broad bodies of literature more efficiently, supporting faster and more comprehensive assessments for climate action.
• Forecasting East African climate patterns - Herbert Misiani from the University of Nairobi shared work using deep learning to predict Kenya’s long rainy season. His presentation showed the value of combining regional climate knowledge with modern modelling approaches.

Debate Shifts Toward Ethics, Equity and Institutional Readiness

A moderated Q&A discussion expanded the conversation beyond technical applications to the broader systems that shape AI use in Africa. Participants raised questions about data governance, legal considerations, intellectual property and the risk of deepening inequalities if AI tools remain accessible only to institutions with strong resources.

Speakers reflected on whether Africa has unique opportunities to lead in AI-driven climate research, given its diverse climates, growing youth population and expanding research networks. Equally, they noted gaps that still limit progress, including uneven computational capacity, limited investment in research infrastructure and the need for stronger collaboration between universities, government and industry.

The discussion also touched on the potential trade-offs of relying heavily on AI models, particularly when transparency, interpretability and community engagement remain essential for climate-related decision-making.

Towards a Practical Toolkit for Emerging Researchers

Closing the session, Timothy Downing (one of the event organisers) reflected on the insights gathered from the discussions through the afternoon. He highlighted a clear appetite among early career researchers for accessible training resources, clearer guidance on ethical practice and opportunities for cross-institutional collaboration.

“My main takeaway from the session in terms of challenges faced by early career researchers in Africa is the lack of massive datasets in Africa that AI needs to be effective. For example, Herbert Misiani 's presentation pointed out that there are actually fewer meteorological stations in Kenya now than there were a couple decades ago- so there's even less data today than there was before with respect to climate,” says Downing.