Sunday, November 22, 2026

Subscribe

PoliticsEdition South AfricaSouth Africa

South Africa seeks international AI oversight within BRICS

South Africa is calling for an international mechanism for independent AI assessment. Pretoria wants stronger safeguards as systems advance faster than the rules.

Editorial meeting in Johannesburg with African journalists, abstract screens and press equipment around an AI dossier.

Regulation is moving more slowly than machines

In the global debate on artificial intelligence, the real dividing line is no longer simply between advocates and the cautious. It now separates those with rules, oversight resources and expertise from those affected by systems designed elsewhere. For South Africa, the stakes are clear: unless governance evolves at the same pace as AI models, countries in the Global South will remain dependent on technical decisions made in a handful of capitals and by a few major companies.

It is in this context that Pretoria is advancing a simple idea with far-reaching implications: establishing an international mechanism for independent scientific assessment of AI, to track capabilities, risks and serious incidents. South African President Cyril Ramaphosa put forward the proposal on 13 September 2026 at a BRICS session in New Delhi, arguing that human governance must advance as quickly as artificial intelligence.

Pretoria joins a momentum already under way in Africa and at the UN

South Africa’s proposal does not emerge from an institutional vacuum. At the continental level, the African Union adopted a Continental Artificial Intelligence Strategy in July 2024, designed to promote the responsible use of these technologies, reduce risks and organise a common approach among member states. In May 2025, the AU even called for the development of a dedicated African AI policy, showing that the continent is seeking to shape the rules of the game globally, not merely accept them.

At international level, the UN established an Independent International Scientific Panel on AI in August 2025, tasked with producing evidence-based assessments. Its first interim report, published in 2026, highlights a gap between the pace of AI capabilities and that of current safeguards. UNESCO, meanwhile, launched a Global Network of AI Supervisory Authorities in 2025 to facilitate exchanges between regulators. In other words, the groundwork is already there. But it remains fragmented.

What South Africa wants: assess, alert and address blind spots

According to the position advanced by Ramaphosa, the proposed BRICS mechanism would rest on several pillars: independent scientific assessments, explicit human oversight, a requirement to report serious incidents, external audits and stronger safeguards as systems become more powerful. This approach reflects the discussions now dominating multilateral forums: how to verify what AI systems actually do, how to document their failures, and how to prevent the race for performance from sweeping aside safety requirements.

For South Africa, the issue is not theoretical. In 2026, the country published a draft national AI policy that prioritised responsible governance, local skills, inclusive innovation and the protection of human rights. But that same year, the government withdrew the draft for revision after criticism over some bibliographical references deemed fictitious. The reversal shows one thing: Pretoria wants to make progress, but is still seeking the right institutional framework. The BRICS discussion therefore also serves to project this search for credibility internationally.

A debate that goes beyond technology to touch on sovereignty

Behind the words “oversight” and “scientific assessment” lies a much broader issue: sovereignty. Who sets safety standards? Who decides that a system has become too powerful, too opaque or too risky? Who has access to data, computing power, laboratories and talent? South Africa, which has one of the continent’s strongest public scientific systems, has for several months stressed the need to develop African AI capabilities rather than depend solely on imported tools.

This perspective also resonates across the rest of southern Africa, where digital transformation is advancing at very different rates across countries, cities and sectors. In public administrations, AI can improve public services. In the informal economy, it can also widen disparities if access to tools, local languages and training remains limited. That is precisely why the AU now links AI to its priorities for development, inclusion and regulation.

What to watch in the coming months

The initiative’s impact will be measured in the forums where it is discussed and then translated into concrete decisions. BRICS is seeking to strengthen its influence on major economic and technological issues. If South Africa’s proposal gains traction, it could help build a more structured South-South approach to AI governance, in line with African positions already advanced by the African Union and with the emerging United Nations framework.

The test will be as much political as technical. Will member states accept independent assessments? Will companies develop audit and reporting mechanisms that are genuinely enforceable? And, above all, will African countries secure a stronger role in setting global standards, even as AI applications are already advancing faster than the rules? That is where the next chapter will be decided: not only in the promise of useful AI, but in the continent’s ability to shape the rules that will govern its future.

Automatic translation. Original French version

The quiz

African institutions: the quiz

6 questions