Glossary

African AI Sovereignty, Defined

A working glossary of the institutions, instruments, and technical concepts that shape Africa's AI agenda.

Sovereign AI
Artificial intelligence systems whose models, training data, compute infrastructure, and governance are owned and controlled within a specific jurisdiction or continent, rather than rented from foreign providers.
African Language Model
A large language model whose training corpus, tokenizer, and evaluation benchmarks include substantial coverage of African languages and African cultural contexts.
Kigali Declaration on AI
The continental agreement signed in April 2025 by 54 African nations establishing a unified African position on artificial intelligence and announcing the $60 billion Africa AI Fund.
Africa AI Fund
The $60 billion continental investment vehicle announced at the Kigali AI summit in April 2025 to finance African AI infrastructure, talent, and applications.
AU Continental AI Strategy
The African Union's official policy framework for AI development across the continent, covering ethics, infrastructure, skills, applications, and continental cooperation.
AfDB AI 10 Billion Initiative
The African Development Bank and UNDP joint initiative targeting $10 billion in AI investment across Africa by 2035.
Smart Africa
The continental policy alliance of 40 African heads of state and 1.4 billion citizens coordinating digital transformation, including AI policy, across member states.
Low-Resource Language
A language for which limited digital text, audio, and labeled data exist, making conventional model training difficult. Most African languages are low-resource in current frontier AI systems.
Accessibility Layer
Kora Lab's product workstream that adapts frontier AI tools for African end users through cultural localization, multilingual interfaces, and affordable delivery.
Sovereign Model Lab
Kora Lab's research workstream that trains, evaluates, and releases African-owned language and multimodal models, datasets, and benchmarks.
Open African Languages Corpus
Kora Lab's open dataset of curated text across African languages, released under permissive licenses for use by researchers and institutions.