Artificial intelligence is everywhere in the talks at UN headquarters during this 81st General Assembly. António Guterres now counts it among the four great challenges of our time, alongside war, inequality and climate change. African countries were quick to take up the issue. South Africa and Somalia both spoke, and they carried the same message: the rules for AI cannot be written without the continent.
South Africa was the first to spell out what it wants. On September 22, the opening day, its Permanent Representative Mxolisi Sizo Nkosi called for the session to end with a statement committing the international community to building « a framework for the global governance of artificial intelligence. » In the diplomat’s view, the way these technologies work, develop and spread needs clear guardrails. Otherwise, the risks will win out.
This stance is part of a broader strategy. At the BRICS summit in New Delhi, Cyril Ramaphosa had already proposed an international mechanism for independent scientific assessment of AI. His plan rests on a few principles: humans must stay in control, incidents must be reported, and developing countries need help to build their own capacity instead of relying on foreign tech giants.
The next day, the debate moved to the Security Council. France, which holds the presidency this month, had organized a high-level meeting chaired by its foreign minister, Jean-Noël Barrot. Canadian researcher Yoshua Bengio sat alongside three industry leaders: Sam Altman of OpenAI, Dario Amodei of Anthropic and Clément Delangue of Hugging Face. With these Silicon Valley figures in the room, Somalia, a non-permanent member of the Council, brought African realities back into focus. It pointed out that the continent still lacks digital infrastructure, energy, data and skills. Without international cooperation to close those gaps, and without a real seat at the table, Africa could end up living under standards designed by the most advanced countries, and existing inequalities would only grow deeper.
Both statements reflect a real concern. Africa is home to nearly a fifth of the world’s population, yet it has almost no weight in building AI models, in computing power or in data centers. Its national AI plans still depend heavily on Google, Microsoft, Nvidia and Meta. African capitals fear a familiar story: rules set elsewhere, then applied to the continent.
The United Nations is trying to respond. During the session, UNDP is due to present a program for Africa that covers AI, data, talent development and access to finance. It is a first step. What remains to be seen is whether the statement South Africa is calling for will materialize, and whether African voices, still few in number, will carry further during the general debate. More of the continent’s heads of state are due to address the Assembly in the coming days.