This paper examines whether existing intellectual property and data governance regimes adequately regulate the cross-border extraction and use of African knowledge in artificial intelligence (AI) systems. As generative AI technologies increasingly depend on large datasets collected across jurisdictions, African cultural expressions, legal materials, indigenous knowledge, languages, educational resources, and creative works are being incorporated into global AI infrastructures, often without meaningful consent, attribution, compensation, or equitable benefitsharing. The paper argues that current copyright and related intellectual property frameworks are insufficient to address the legal and structural inequalities created by transnational AI data practices, and it proposes a reform-oriented framework built on transparency, cross-border benefitsharing, recognition of collective knowledge interests, and equitable participation in AI governance.