Ethical AI NIASET

Ethical AI for Africa

According to UNESCO, roughly 26 languages disappear each year in the world. Many cultures in Africa are sustained by oral tradition and when a language ceases to be spoken, part of the culture dies as well. At the SankofAI Project, we are acutely aware of the urgency underlying our mission. In a decade, it might already be too late to achieve a meaningful representation of African traditional culture in AI models. Currently, major AI companies concentrate on training their models in domains such as finance, mathematics, physics, language translation, medical diagnosis and software coding. These are all applications that promise significant future profits. The ritual scarification practices among the Moru people of South Sudan or the hairstyles of the Himba people in Namibia offer little pecuniary advantage. However, it is the right thing to do. Ethical AI must look beyond mere financial gain and invest equally in efforts that will also enrich humanity. The SankofAI Project and NIASET: Africa's foundational dataset are a good start.

SankofAI / NIASET Ethical Framework

“Where ancestral wisdom becomes AI’s conscience.”

CORE PRINCIPLES

1. Cultural Sovereignty

The right of African peoples and communities to define, share, and protect their cultural knowledge and traditions in the digital age.

SankofAI recognizes indigenous and traditional knowledge as intellectual property belonging to the communities that created it.

Any data used—visual, oral, ritual, linguistic—must be approached with respect for its meaning, origin, and custodians.

2. Representation Justice

Ensuring that AI models include Africans not as monoliths or stereotypes, but in all their ethnic, historical, and cultural complexity.

We actively counteract erasure, tokenism, and distortion by curating data that accurately reflects Africa’s multilingual, multitribal, and multigenerational identity.

Inclusion is not passive—it's intentional, guided by experts, and accountable to the communities represented.

3. Participatory Design

“Nothing about us without us.”

Community knowledge holders, artists, historians, linguists, and spiritual leaders are not sources of data—they are co-creators.

SankofAI commits to consultation, consent, and compensation at every stage of the data lifecycle.

4. Transparent Governance

Ethical oversight that is open, interdisciplinary, and accountable.

An independent Cultural and Ethical Advisory Council will guide project decisions.

All data sources, methodologies, and use cases will be publicly documented to ensure auditable AI practices.

5. Minimizing Harm, Maximizing Dignity

Sacred knowledge must not become speculative content.

Some traditions are not meant to be digitized or reimagined through AI. SankofAI respects the right to silence.

AI outputs will be designed with cultural red lines—boundaries that must not be crossed, such as sacred rituals, funerary practices, or initiations.

Vision for the Future

NIASET is more than an AI data collection project. It is a cultural preservation movement and a technological decolonization effort. By embedding ancestral knowledge ethically into future systems, we help ensure that:

Africa is not remembered through the eyes of others but understood through the voice of its own people.

We believe the digital future belongs to everyone—but only if it is built ethically, inclusively, and respectfully.

The Imperative of African Narratives in Shaping Global AI

The pervasive influence of Artificial Intelligence (AI) is undeniable, shaping virtually every facet of modern life, from healthcare and education to economic opportunity and social interaction. However, a critical void exists within the global discourse on AI narratives: the profound absence or marginalization of perspectives emanating from Africa. This omission is not merely an oversight; it carries a significant "cultural cost" and holds profound implications for how AI systems are designed, deployed, and regulated, particularly within the African continent itself.

The scarcity of data reflecting Africa's rich tapestry of life philosophies, such as Ubuntu, alongside its diverse thought processes, spiritual beliefs, traditions, and customs, creates a dangerous vacuum in AI development. This analysis will illuminate the inherent dangers arising from this exclusion and underscore the immense, untapped benefits of integrating African perspectives. It also argues for the SankofAI Project as a vital and timely solution.

The development of AI without adequate representation of African narratives presents a multifaceted array of risks, ranging from the perpetuation of systemic biases to the erosion of cultural heritage and the exacerbation of digital inequalities.

AI systems are dependent on the data upon which they are trained; consequently, they are "only as good as the data employed". This fundamental characteristic means that if these datasets reflect historical inequalities and prejudices, the AI models inevitably transfer and amplify these biases, perpetuating existing disparities. This phenomenon can lead to what has been termed "algorithmic apartheid," a digital mirroring of historical injustices. For these reasons, we are building NIASET.

Click here to download a pdf of the full analysis

SankofAI | Ethical AI