AI: Eswatini opens an African center in Mbabane to evaluate and certify systems
Eswatini has established a center of excellence for artificial intelligence safety and standardization, launched on October 27, 2026, in Mbabane by the Eswatini Standards Authority (ESWASA) and the organization AICE Africa. The stated objective is to build the capacity in Africa to evaluate, govern, and certify AI systems, according to Capital FM, which reported the announcement.
The issue is less technical than it is political. The AI tools deployed across the continent—in banks, hospitals, and government offices—are essentially audited elsewhere, according to standards designed elsewhere. The CEO of AICE Africa, John Kamara, summarized the center's ambition: "AI sovereignty requires institutions capable of evaluating" the systems we depend on, he stated, explaining that he wants to reduce African dependence on external institutions.
A laboratory designed for African languages
The project includes a research and testing laboratory tasked with evaluating AI systems against international standards, while taking into account African languages, datasets, infrastructure, and the continent's social conditions. The center intends to rely on two frameworks: the ISO/IEC 42001 standard, which sets requirements for AI management systems, and the ISO/IEC 23894 standard, dedicated to risk management. This laboratory remains in the project stage.
The division of roles is established: ESWASA provides institutional leadership, standardization expertise, and national coordination; AICE Africa provides expertise in AI governance, technical implementation, and skills transfer. The executive director of ESWASA, Ncamiso Mhlanga, justifies the approach simply: "as AI becomes part of the functioning" of businesses and public services, standardization bodies must know how to evaluate it.
In its first phase, the center will focus on the internal transformation of ESWASA before opening up to professional training, business services, and regional cooperation. A platform called EswasaOne is under development to integrate normative information, certification, and inspection processes. Training is planned for auditors, regulators, and technical staff, with the support of universities.
Caution remains the order of the day on one point: the issuance of certifications will be "explored," but will depend on validated evaluation methods, qualified personnel, and accreditation agreements.
The initiative comes, notes Capital FM, at a time when governments and companies are integrating AI into public services and commercial activities, which raises questions about data protection, reliability, accountability, and risks related to automated decisions. The partners have announced a summit on AI safety and standards in Mbabane in July 2027 to expand regional cooperation.
This article was written in French with the help of artificial intelligence from the sources cited below, then translated from French automatically. Read the original.
