African Observatory on Responsible AI

Congratulations to Msa Mesatywa on completing the Economics of Artificial Intelligence in Africa course! We couldn't agree more, AI is indeed reshaping markets, labour, and opportunity across Africa, and the economists, policymakers, and entrepreneurs who understand that shift will be the ones who shape it. Your reflections speak to the heart of what this programme was designed to do, build the understanding that ethical AI adoption, financial inclusion, and equitable policy frameworks are deeply interconnected. In partnership with the University of the Witwatersrand and Google this course aims to build a network of African practitioners, researchers, and policymakers who can engage AI's economic dimensions with the depth and nuance the continent deserves.

New research: Predictable Confabulations: Factual Recall by LLMs Scales with Model Size and Topic Frequency New research co-authored by Samuel T. Segun, PhD, Dr. Matthew Smith, Prof Jonathan Shock, Dr. Olatunji Iyiola Emmanuel (李白) and Prof. Tegawendé F. Bissyandé. evaluated 38 AI models across more than 8,900 scholarly references spanning 24 topics. The findings show that factual recall quality follows a measurable pattern determined by model size and how well a topic is represented in training data. The development relevance is stark. Topics like climate change, with over a million associated scholarly works, returned high-quality results even from mid-sized models. Topics like school dropout prevention in rural areas, returned poor results even from the largest models tested. The authors estimate that achieving comparable reliability on such a low-frequency topic would require a model roughly 30 times larger than anything currently available. For AI deployed in health, agriculture, education, and governance contexts across lower-and middle-income countries, where locally relevant training data is frequently limited, this is a governance challenge. Read the full paper and access additional resources: https://lnkd.in/eymP6A7N

In our latest Women in Focus spotlight, we feature Dr. Chinasa T. Okolo, Ph.D.. Okolo, Founder and Scientific Director of Technecultura, a nonprofit research institute and policy incubator advancing global equity in AI. Chinasa works at the intersection of AI governance, policy, and inclusion, with a particular focus on AI safety for Global Majority communities, the sociocultural limitations of AI systems, and the geopolitical implications of emerging technology for the African continent. In her interview, she reflects on how her Nigerian-American background has shaped her commitment to centering the African continent in global AI governance conversations. She also speaks openly about the gaps she sees in African AI policymaking, what inclusive policy development actually requires, and the advice she wishes she had received starting out. 📖 Read the full interview: https://lnkd.in/dhChZtvd

The gap between global AI commitments and reality Countries have signed AI declarations, adopted UNESCO recommendations, endorsed the Bletchley commitments and participated in the Global Digital Compact. But commitments made at summits don't automatically become the frameworks and enforcement mechanisms that protect people. That's the gap GIRAI was built to measure. The 2nd edition of the Global Index on Responsible AI, launching in July 2026, tracks whether countries are turning international AI commitments into real governance. Not just the policies but the structures, the oversight, the accountability. Because the distance between "we agreed to this" and "we did this" is exactly where rights get lost. Join our mailing list to be among the first to receive the 2026 edition: https://lnkd.in/d4ifHfVn International Development Research Centre (IDRC) Foreign, Commonwealth and Development Office Inter-American Development Bank Global Affairs Canada | Affaires mondiales Canada Artificial Intelligence for Development (AI4D) Local Development Research Institute LIRNEasia ILDA #AIGovernance #GIRAI2026 #ResponsibleAI #AIPolicy #HumanRights

This Workers' Day, we're thinking about the workers AI is least likely to protect. The women, low-income earners, and caregivers who perform the work that holds our societies together, yet remain most at risk of being left behind by the digital economy. Globally, women perform up to ten times more unpaid care work than men. They are most likely to be automated out of jobs, and least likely to benefit from new AI-driven opportunities. On the African continent, these dynamics are made worse by race, class, and histories of structural exclusion. Jess C.'s piece for The Algorithmic Review confronts this reality head-on. She argues that inclusion in the digital economy is not enough if the systems being built reproduce the same hierarchies we're trying to dismantle. Her call is to centre care to be seen as a value, in how we design AI, measure work, and imagine the future of labour. Read the full piece: https://lnkd.in/dygD2he4 #WorkersDay #LabourDay #CareWork

𝗪𝗵𝗮𝘁 𝗱𝗼𝗲𝘀 𝗔𝗜 𝗴𝗼𝘃𝗲𝗿𝗻𝗮𝗻𝗰𝗲 𝗿𝗲𝗮𝗹𝗹𝘆 𝗹𝗼𝗼𝗸 𝗹𝗶𝗸𝗲 𝗮𝗰𝗿𝗼𝘀𝘀 𝗔𝗳𝗿𝗶𝗰𝗮? With over 1000 entries spanning national AI strategies, policies, and governance frameworks from across the continent, the Africa AI Policy Tool is the most comprehensive resource of its kind for understanding how Africa is approaching AI governance. The tool is designed to support peer learning, comparative analysis, and evidence-informed policymaking because understanding how African countries are governing AI is the first step toward governing it well. Explore the tool and see the diversity of approaches African countries are taking to govern AI responsibly. 🔗 https://lnkd.in/dFTEWucB Policy Action Network (PAN) International Development Research Centre (IDRC) Foreign, Commonwealth and Development Office

Call for Applications ends today, April 20, 2026. Apply for LUNE THREE - Shaping AI from African Contexts: https://lnkd.in/eGtrSU_E #LUNEThree #AIinAfrica #ResearchRound

As we reflect on what we must preserve this World Heritage Day, we celebrate the ways of knowing, gathering, and deliberating that communities have practised for generations. In this piece, Selamawit Engida Abdella traces the history of AI starting from the 1956 coining of the term. She explains that the founding context still shapes who defines AI, who funds it, and who benefits. Africa receives less than 1% of global AI investment, yet the continent's land, labour, and data are increasingly being drawn into AI-driven development. This piece is part of the "Ideas from the Palaver Tree" collection, that centres African voices in the responsible AI conversation. The collection features 6 articles and we invite you to learn from, connect with, and even challenge the ideas presented in this collection. 🔗 Link in comments.