Abstract
This research-in-progress paper examines how people in low‑income countries (LICs) use generative AI (GenAI) in practice. Using country-level data from Anthropic’s Claude Economic Index, we characterize usage across 13 LICs by collaboration pattern, task, and request complexity. We consolidate patterns into automation and augmentation. All countries are automation‑leaning; the most balanced are Maldives, Togo, and Uganda, while Burundi, Chad, Gambia, Liberia, and Sierra Leone show exclusively automation-focused use. Task data (available for five countries) indicate a dominance of software maintenance/upgrades, followed by new development, with education support prominent in Mozambique. Coverage is highest at complexity level‑2 (53.19% classified), lower at level‑1 (36.48%), and minimal at level‑0 (8.83%, Mozambique only). We discussed the empirical findings and practical implications.
Recommended Citation
Ning, Xue; Yu, Yixiu; and van Solt, Michelle, "Usage of Generative Artificial Intelligence (GenAI) by People in Low-Income Countries" (2025). GlobDev 2025. 9.
https://aisel.aisnet.org/globdev2025/9