From Private Database to Public Resource to a Testing Ground for AI Methods: The Islam West Africa Collection (IWAC)

This article traces the evolution of the Islam West Africa Collection (IWAC) from a private Zotero database into an open-access digital resource hosted at ZMO, and its emerging role as a testing ground for digital humanities and AI methods. IWAC now contains over 14,000 items — archival documents, newspaper articles, Islamic publications, audio and video recordings, and photographs on Islam and Muslims in Benin, Burkina Faso, Côte d'Ivoire, Niger, Nigeria, and Togo. The article discusses the project's grounding in open science and FAIR principles, the copyright negotiations that enabled public access, and its focus on newspapers and Islamic print culture as sources that complement manuscript-centred digitisation efforts. It then examines three AI applications: multimodal LLMs for OCR of complex newspaper layouts, automated sentiment analysis of media representations of Islam, and a retrieval-augmented generation (RAG) chatbot for semantic search across the corpus. These AI-assisted approaches are presented as practical strategies for sustaining a single-researcher project at scale, with broader implications for digital humanities in African studies.