AUDA-NEPAD Launches ATLAS Umoja AI Policy Workstream

The African Union Development Agency–NEPAD has launched the ATLAS Umoja AI policy workstream, adding a continental governance track to a broader effort to build artificial intelligence that reflects African languages, cultures and public priorities.

The workstream was announced after an AUDA-NEPAD Presidential Roundtable attended by GSMA Chief Regulatory Officer John Giusti. The initiative is intended to bring questions about AI governance into discussions involving African institutions and government stakeholders, rather than leaving those decisions to technology providers based outside the continent.

The language gap is one reason the policy debate matters. Giusti said fewer than 2% of African languages are meaningfully supported by current AI models. That is not only a model-development problem. It affects whether people can use digital services, search for information or interact with public systems in languages they understand.

ATLAS Umoja is built around a wider open platform that brings together governments, industry, researchers and innovators to develop benchmarks, models and communities for African-language AI. The platform’s stated scope includes all 54 African countries, with country-led models and datasets released for developers to evaluate and build on.

The policy workstream should help answer practical questions that technical projects alone cannot settle. Who owns language data? What standards should determine whether a model is safe or accurate? How can countries share infrastructure without giving up control over national priorities? And how should governments measure whether local-language AI improves access rather than simply adding another layer of technology?

The launch is therefore more than a new AI programme label. It signals that language, data ownership and governance are becoming part of the continent’s AI infrastructure. The next test will be whether the workstream produces common principles that can guide procurement, model evaluation and public-sector deployment across different national systems.

Source: TechReview Africa. Additional platform context: ATLAS Umoja AI.