From Storytelling to Ethical AI

African evaluation scholars such as Bagele Chilisa and colleagues (2016) advance a relational evaluation paradigm grounded in community, spirituality, and interconnectedness, offering an alternative to dominant Western models. Rooted in philosophies such as Ubuntu, this approach emphasizes participatory, culturally embedded methods and forms part of broader efforts to decolonise and indigenise evaluation practice by centering African epistemologies and community ownership of knowledge.
This session presents findings from the seedling phase of a multi-country research initiative led by the Africa Gender and Development Evaluators Network (AGDEN) under the EvalIndigenous “EI-Indexing Seeding Grants.” The project explores indigenous knowledge systems in Lesotho and Nigeria through three small-scale studies, using mixed methods alongside digital artefacts, including short video presentations that documents storytelling, cultural practices, and community perspectives. These video-based research presentation serve as both data and dissemination, preserving voice, and context often lost in text-based analysis.

Across the studies, findings reveal that indigenous knowledge expressed through Basotho cultural practices, the endangered Ikom Monoliths of Nigeria, and Berom indigenous business systems is deeply relational and transmitted through oral and community-based processes. These systems hold critical insights relevant to global challenges such as climate resilience, inequality, gender and yet remain underrepresented in dominant evaluation frameworks.

Engaging directly with AI-enhanced evaluation, the session asks whether artificial intelligence can meaningfully capture such knowledge. While AI may help surface marginalized perspectives and detect patterns invisible to traditional approaches, it also introduces risks, including decontextualization, algorithmic bias, and the exclusion of non-textual, culturally embedded knowledge. These tensions raise critical questions about trust, representation, and epistemic justice in evaluation.

The germinating phase of the project responds by exploring how AI can be reimagined within African evaluation contexts to address complexity more ethically and effectively. It emphasizes safeguards such as cultural integrity, community consent, and data sovereignty, while reaffirming the essential role of human evaluators in contextual interpretation and meaning-making.