Senegalese AI expert Seydina Moussa Ndiaye, speaking in an on-the-record interview with the United Nations, warned that foreign technology companies risk driving a new form of colonization across Africa by using African data to power artificial intelligence systems, with much of the value flowing abroad. His comments arrive at a moment when the UN is actively building governance frameworks for AI, including the Global Digital Compact and a high-level advisory report on equitable AI oversight. The tension at the center of this debate is straightforward: African societies generate valuable digital data, but many of the companies training AI models are headquartered elsewhere, raising hard questions about who profits and who loses control.
What is verified so far
Ndiaye’s central claim is blunt. In the interview conducted by the UN Information Service in Nairobi, the Senegalese expert stated that foreign companies could feed on African data and, in doing so, replicate historical patterns of resource extraction under a digital banner. He used the phrase “digital colonization” to describe a scenario in which AI-driven data flows benefit external firms while African nations see little return. The framing is deliberate: it draws a direct line between colonial-era commodity extraction and the modern harvesting of mobile usage records, cultural archives, and behavioral datasets across the continent.
This warning does not exist in isolation. The UN has been building institutional infrastructure around these exact concerns. The Global Digital Compact commits member states to inclusive digital cooperation and explicitly addresses reducing digital divides. The compact’s language on sovereignty and equitable participation is directly relevant to Ndiaye’s argument. If African governments lack the capacity or legal tools to regulate how their citizens’ data is collected, processed, and monetized by foreign AI firms, the compact’s goals remain aspirational rather than operational.
Separately, on September 19, 2024, UN Secretary-General António Guterres delivered a video message at the launch of the final report from the High-Level Advisory Body on Artificial Intelligence, titled “Governing AI for Humanity.” Guterres stressed the urgency of equitable AI governance and bridging what the UN calls “AI divides,” gaps in access, capacity, and influence that risk leaving lower-income nations on the wrong side of the technology revolution. The advisory body’s recommendations, as reported in coverage from the Associated Press, urge the United Nations to lay the foundations for global AI governance, a clear signal that the institution views the current regulatory vacuum as a serious risk.
Taken together, these developments show a coordinated push within the UN system to name and confront the power imbalance in global AI development. Ndiaye’s interview, the Global Digital Compact, and the advisory body’s report all point toward the same conclusion: without enforceable rules, AI will deepen existing inequalities between data-rich but governance-poor regions and the handful of nations and corporations that dominate AI research and deployment.
What remains uncertain
Several critical gaps remain in the evidence base. No publicly available UN research or dataset quantifies how much African data is currently being extracted by foreign AI companies, or which specific firms are doing the extracting. Ndiaye’s warning is framed as a forward-looking risk rather than a documented pattern with hard numbers attached. This matters because the difference between a plausible warning and a verified crisis is measurable evidence, and that evidence has not yet surfaced in the institutional record.
Equally unclear is how African governments and regional bodies plan to respond. The African Union, individual national regulators, and regional economic communities have not issued specific AI data sovereignty policies that are cited in the available UN reporting. Without knowing what legal tools African states currently have, or are actively building, it is difficult to assess whether the “digital colonization” risk is imminent or still theoretical. Some African nations have begun drafting data protection laws modeled on the European Union’s General Data Protection Regulation, but the scope and enforcement capacity of these efforts remain uneven and poorly documented in the sources reviewed here.
The full text of the “Governing AI for Humanity” report has not been independently analyzed in the available reporting beyond the AP’s summary and Guterres’ video message. That means the specific recommendations the advisory body makes about Africa, if any, are not yet confirmed in detail. It is possible the report addresses continent-specific risks directly; it is also possible that its recommendations are global in scope and do not single out Africa’s situation. Until the full document is reviewed, this remains an open question.
Ndiaye’s own credentials and institutional role are described only through UN news channels. No independent biographical profile or academic record was available in the reporting block to verify his specific advisory role or research background beyond what the UN interview states. This does not discredit his claims, but it does mean readers should treat his authority as institutionally vouched rather than independently confirmed. The UN often highlights regional experts to illustrate emerging issues, but the depth of their academic or policy influence can vary widely.
Another uncertainty concerns enforcement. Even if global standards emerge from UN processes, it is not yet clear how binding they will be on powerful technology firms headquartered outside Africa. Many of these companies operate across multiple jurisdictions and can shift data-processing operations to countries with looser regulations. Without clear mechanisms to monitor compliance or impose penalties, African states could find themselves with strong principles on paper but limited leverage in practice.
How to read the evidence
The strongest evidence in this story comes from primary UN sources: the on-the-record interview with Ndiaye, the official Global Digital Compact page, and the Secretary-General’s statement launching the advisory body’s final report. These are institutional documents that reflect official UN positions and on-the-record statements from named individuals. They carry significant weight because they represent what the UN system is willing to say publicly and attach its credibility to.
The AP’s coverage of the advisory body’s report release adds independent editorial verification. It confirms the event occurred, summarizes the report’s key recommendations, and provides a journalistic filter that is separate from the UN’s own communications apparatus. This is useful for readers trying to distinguish between what the UN says about itself and what outside observers confirm.
What the evidence does not yet provide is the quantitative backbone that would turn Ndiaye’s warning into a fully substantiated diagnosis. There are no public figures on the volume of African data used in training major AI models, the proportion of AI investments returning to African economies, or the number of African-owned AI firms participating in global value chains. In the absence of such metrics, the “digital colonization” metaphor remains a powerful but largely qualitative frame.
Readers should therefore interpret Ndiaye’s comments as an early-warning signal grounded in structural realities rather than as proof of an ongoing, measured exploitation. The structural realities are well documented: Africa’s rapidly growing population of mobile phone users, expanding internet connectivity, and rich linguistic and cultural diversity all generate data that is attractive to AI developers. At the same time, the continent hosts relatively few large-scale data centers, AI research labs, or venture-backed technology companies compared with North America, Europe, or parts of Asia. This imbalance between data generation and value capture is the backdrop for the colonization analogy.
It is also important to situate the UN’s role. The organization can set norms, convene experts, and encourage capacity-building, but it does not directly regulate multinational technology firms. Its influence depends on whether member states translate high-level compacts and advisory reports into domestic legislation and regional agreements. Efforts to build technical and regulatory capacity within African institutions may be supported through UN programs, but such initiatives operate on long timelines and often with constrained resources.
For now, the evidence supports three cautious conclusions. First, the risk Ndiaye describes is plausible and aligned with broader concerns about global AI inequality, but it is not yet backed by detailed public data on African-specific exploitation. Second, the UN is moving to articulate principles and governance frameworks that could, in theory, help prevent such exploitation, yet the effectiveness of these tools will depend on political will and enforcement capacity that are not guaranteed. Third, African voices like Ndiaye’s are beginning to shape the narrative around AI governance, but much of that influence still flows through UN-managed platforms rather than independent academic or civil-society research.
As debates over AI governance intensify, the question for African policymakers, technologists, and citizens is whether they can convert early warnings into concrete safeguards: robust data protection laws, regional standards on cross-border data flows, investment in local AI capacity, and stronger participation in global rule-making forums. Until such measures are in place and their impact can be measured, “digital colonization” will remain a contested but resonant term for a future that many in Africa are determined to avoid, even as the world’s most powerful AI systems continue to hunger for the data their societies produce.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
