African leaders are demanding a greater role in setting global standards for artificial intelligence. Their call comes amid growing concerns about AI misuse and malfunction, with incidents reported across the continent. This demand was emphasized last week at the United Nations, where leaders highlighted the urgency of African involvement in shaping AI governance.
At a recent U.N. Security Council meeting, Lewis Garseedah Brown II, permanent representative of Liberia, emphasized the need for Africa to be an equal partner in shaping the future of AI. “Africa must be equal co-architects in determining the standards, ethics, and architectures of this technological era”, he said. His statement reflects a broader sentiment among African leaders who see AI governance as a critical area for continental autonomy.
AI incidents in Africa
Incidents of AI misuse and malfunction are already occurring in Africa. In Kenya, a predictive machine learning algorithm for setting health insurance contributions has increased costs for poor households. Scammers in West Africa have used AI-generated content to blackmail victims in extortion schemes. The militant group Boko Haram has also employed AI to organize attacks and design weapons after bypassing safeguards. These cases show the immediate risks AI poses in diverse African contexts.
According to Jonathan Shock, associate professor of mathematics at the University of Cape Town, there is a lack of discussion about AI safety. “There’s essentially no real discussion about AI safety. So, there’s a real vacuum”, he said. Shock warns that a catastrophic incident might be necessary for governments to prioritize AI safety. “The pessimistic view is we don’t stop it in time”, he added, highlighting the potential consequences of inaction.
AI adoption and risks in Africa
Despite these risks, African companies are rapidly adopting AI. A survey by PricewaterhouseCoopers found that out of 85 large African companies surveyed this year, over 80% were running AI pilots. However, many lack formal structures to oversee safety, leaving them vulnerable to unforeseen challenges. This rapid adoption without adequate safeguards raises concerns about long-term implications.
Mélanie Keïta, CEO of Nairobi-based financing company Melanin Kapital, warns about the risks of relying on foreign AI models. “We might become a product of other foreign corporations that own all our data as opposed to really being autonomous and free”, she said. Her company has tested both American and Chinese AI models, each presenting unique risks, further complicating the decision-making process for African businesses.
The risk associated with AI models depends on how easily they can be shut down if issues arise. Closed-weight models like ChatGPT and Claude are centrally controlled, while open-weight models like DeepSeek and Kimi can be downloaded and modified by users, making remote shutdown more challenging.
As African governments integrate AI into sensitive areas like healthcare and education, the need for robust safety evaluations becomes critical. Gathoni Ireri, a researcher at the Nairobi-based Ilina Program, advocates for “Africa-centric” safety evaluations that consider each country’s unique economic and social contexts. She emphasizes that harm should be assessed relative to the size of each country’s economy, as incidents that might seem minor elsewhere could have devastating effects in Africa.
The need for African-led AI governance
Currently, fewer than half of African countries have AI policies or strategies in place. Kenya is the only African nation participating in an international AI safety network, which focuses on global risks rather than Africa-specific ones. This limited participation shows the continent’s struggle to influence global AI governance frameworks.
African nations must develop the capacity to conduct their own safety evaluations, according to Ali Mohamed Omar, Somalia’s state minister for foreign affairs. “Africa cannot remain merely a market for systems developed elsewhere, a source of raw data, or a testing ground for a technology it did not help govern”, he stated.