African Agency in the Digital Turn

In the AI Story

The argument is partly empirical and partly normative. Empirically, there is already a vibrant African AI ecosystem that the metropolitan discourse largely ignores. Masakhane, the community-driven machine translation effort for African languages, has produced state-of-the-art models for dozens of languages through participatory research methods that contrast sharply with the extractive practices of major labs. Lelapa AI in South Africa builds models specifically for African contexts. The Lacuna Fund supports African researchers developing datasets for underrepresented problems. The African Content Moderation Workers' Union is organizing the labor side. These efforts do not require metropolitan validation to exist, and they are producing intellectual and technical artifacts that the metropolitan center could learn from if it were willing.

Normatively, the argument is that the continent's intellectual traditions — oral memory cultures, collective improvisation in music and language, the negotiation with and through colonial languages, the experience of building parallel institutions under hostile conditions — are directly relevant to the problems the AI age poses. The question of how communities retain knowledge against its appropriation is not a new question for African thinkers. The question of how to build under conditions of infrastructural dependency is not new either. The question of how to improvise when the system does not work as designed is a competency the continent has developed for centuries.

The framework cuts against the tendency of both tech optimism and tech criticism to treat Africa as a passive screen onto which the hopes or fears of metropolitan commentators are projected. The developer in Lagos is not a hypothetical figure in an argument about whether AI democratizes creativity. She is a specific person, with specific intellectual lineages, making specific choices about which tools to use, which to build, and which to refuse. What she does with AI will be determined by her agency, operating under the constraints Mbembe's other concepts describe, not by the prior commitments of commentators in San Francisco or Berlin.

The generative dimension of Mbembe's work has sometimes been underread. Critics have treated him as a purely diagnostic thinker, useful for exposing colonial continuities but not for building alternatives. This is a misreading. Mbembe's later work — especially Out of the Dark Night — is sustained in a more constructive register, proposing what he calls a planetary ethics and a politics of the common that requires contributions from the margins, not at their expense.

Origin

The argument runs through the later stages of Mbembe's work and connects to a broader African intellectual tradition including Achille Mbembe himself, Souleymane Bachir Diagne, Felwine Sarr, Sylvia Tamale, and the various communities engaged in decolonial technology practice on the continent.

Key Ideas

Between triumph and catastrophe. African agency cannot be seen through either the celebratory or the tragic frame alone.

Existing ecosystem. African AI work is already producing significant technical and intellectual artifacts, largely invisible to metropolitan discourse.

Traditions as resources. Oral creativity, improvisation, collective intelligence, and the experience of parallel institution-building are resources for the AI age.

Refusal is a form of agency. The choice to not build on certain platforms, not participate in certain extractions, is a form of agency as real as adoption.

Generative, not only diagnostic. The framework points toward what African engagement with AI can produce, not only what it must resist.