Digital Colonialism and Algorithmic Power: Data Sovereignty and AI Governance in the Global South

Africa's integration into the global artificial intelligence ecosystem has intensified concerns about digital colonialism. The continent increasingly supplies data, labour and digital markets for AI systems while exercising limited control over the infrastructures, standards and rules that govern them. This article argues that the resulting dependency is reproduced through four mechanisms: data extraction and value capture, infrastructural dependency, asymmetric rule-making and algorithmic power. It examines African legal and policy responses, including data protection laws, the Malabo Convention, the AfCFTA Digital Trade Protocol and the African Continental AI Strategy, against influential external frameworks such as the OECD AI Principles and the EU AI Act. India and Brazil are used as bounded comparative cases to test whether stronger state capacity or rights-based legal frameworks overcome algorithmic dependency. The article concludes that they do not eliminate dependency but reconfigure it. It therefore proposes the Algorithmic Dependency and Data Sovereignty Model (ADDSM) as a diagnostic tool for sequencing governance, infrastructure, economic and enforcement reforms.