Digital Humanities and Artificial Intelligence in African Studies

Digital humanities (DH) and artificial intelligence (AI) are changing how knowledge is produced in African studies. They make new kinds of analysis possible, and new narratives with them: reimagining cultural heritage, widening access to diverse sources, amplifying marginalised voices. As the field approaches an AI-driven “tipping point”, the project explores methods for designing, evaluating and sustaining digital infrastructures. These systems need to be grounded in African epistemologies rather than retrofitted from Western models.

These questions run through a collaborative initiative: two workshops in 2026 (Madore and Hiribarren, 2026) (Madore et al., 2026) and a co-edited volume, Digital Humanities and Artificial Intelligence in African Studies, to be published by Bielefeld University Press. The project brings together scholars and practitioners from Africa, Europe and beyond, and opens dialogue between the Global North and South. Three themes organise the work: transforming research methods through computational tools, building sustainable infrastructures in resource-constrained settings, and centring African knowledge systems in digital design. The move is from policy governance about AI to “practising with AI”, and towards the ethical development of DH across the continent.

My contribution addresses the practical “problem of scale” in African digital collections. Tracing my own route from “digital hoarder” to “historian-programmer”, I explore how far agentic coding lowers the barrier to entry: the researcher specifies a problem in natural language, and an AI agent writes, runs and debugs the code. Pipelines built on large language models (LLMs) take over labour-intensive work such as optical character recognition (OCR) and named entity recognition (NER). These Python-based workflows open up vast, under-resourced collections. The same approach turns a static archive into interactive visualisations: networks of who appears alongside whom, word frequencies traced year by year, a semantic map that places 12,000 articles by textual similarity. Most link back to the documents behind them, so readers can test their own questions against the sources.

I prototype these methods in the Islam West Africa Collection (IWAC), whose 30-million-word corpus tests the limits of AI in historical research (Nobel-Dilaty and Madore, 2025). Working from Moretti’s (2000) “distant reading”, I treat AI as a “co-intelligence” (Mollick, 2024) for corpora too large to read in their entirety. AI-assisted sentiment analysis, for instance, examines how Islam and Muslims are represented in Francophone West African newspapers, with the results visualised in an interactive dashboard. Human judgement still decides what the patterns mean; AI offers a workable way of handling the volume of digital material African history now generates.