<p>Africa’s AI-enabled digital labour platforms operate within severe structural asymmetries including limited electricity access, minimal African representation in training data, unequal digital skills distribution and underdeveloped governance frameworks. Generative AI systems—large language models, agentic orchestration and automated quality control—intensify these challenges while creating new opportunities. Current research inadequately addresses African socioeconomic realities, positions stakeholders as passive adopters, and lacks integration between platform governance theories and decent work outcomes. This paper advances a conceptual framework theorizing how Ubuntu-aligned governance mechanisms moderate relationships between AI platform mechanisms and worker outcomes, contributing to digital platform sovereignty. The framework integrates structural asymmetries analysis with five AI mechanisms (three legacy, two Gen-AI-specific), Ubuntu governance levers operationalizing relational ethics, seven decent work dimensions and boundary conditions shaping applicability. Six testable propositions specify moderation effects enabling empirical investigation through publicly available data, worker surveys and participatory methods.</p>