South Africa Calls for Global AI Guardrails at UN General Assembly

South Africa is taking its AI governance argument to the United Nations, calling for a global framework that can guide how artificial intelligence is developed and deployed.

The position was set out by Mxolisi Sizo Nkosi, South Africa’s permanent representative to the UN in New York, ahead of the 81st General Assembly. Speaking to SAnews, Nkosi said the country wants the international discussion to move beyond the promise of new technology and toward rules that can manage its risks.

That distinction matters. AI systems are being developed and released across borders, while the consequences of their use often fall within national systems that have very different regulatory capacity. A model trained in one jurisdiction can influence public services, labour markets or information environments in another without a shared understanding of who is responsible when things go wrong.

South Africa’s proposed outcome is a commitment by countries to work toward a global governance framework. Nkosi linked the issue to concerns about increasingly powerful systems, uncontrolled model behaviour and the need for safeguards that do not depend solely on voluntary action by technology companies.

The argument also places AI within Pretoria’s broader foreign-policy agenda. South Africa is expected to raise the subject alongside peace and security, climate change, reform of international institutions and inequality. In that framing, AI governance is not a specialist technology issue; it is part of the debate over how global systems distribute power and opportunity.

For African governments, the question is particularly consequential. Rules written without meaningful participation from developing countries could leave the continent adapting to standards it had little role in shaping. Yet a global framework will only be credible if it can be translated into national institutions that have the expertise and resources to implement it.

South Africa’s intervention therefore carries two tests. The first is diplomatic: whether countries can agree on common principles as AI capabilities accelerate. The second is practical: whether those principles improve accountability, protect rights and widen access to the scientific capacity needed to evaluate advanced systems.

The UN discussion will not settle those questions on its own. But it signals that South Africa wants Africa’s voice to be present while the international rules of the AI era are still being negotiated.

Source: SAnews.