Artificial Intelligence (AI) and the New Colonialism of Climate Data in the Global South

The paper examines how AI tools for climate adaptation, primarily designed and governed by global North institutions, impose top-down risk representations that exclude the lived experiences and indigenous knowledge of global South communities. Using Cape Town's informal settlements as a case study, it illustrates how algorithmic abstraction can lead to ineffective policies and epistemic injustice. The author calls for a paradigm shift toward data sovereignty, participatory design, and epistemic pluralism to prevent 'smart' adaptation from replicating historical colonial patterns of exclusion.

• AI-driven climate data systems are predominantly designed, funded, and governed by global North institutions and corporations, creating a power asymmetry where Northern epistemologies and capital logics shape climate governance.
• The use of standardised metrics and large-scale datasets in climate AI often results in 'epistemic injustice,' where the quantifiable knowledge of distant institutions is privileged over the situated, lived experiences and indigenous knowledge of Southern communities.
• In Cape Town's informal settlements, AI flood-risk tools have demonstrated a misalignment with local realities, such as designating areas as 'non-viable' for upgrading while ignoring existing community-led drainage innovations.
• Global South countries face significant policy gaps in climate AI governance, including a lack of ethical regulatory frameworks, an overreliance on imported Northern systems, and insufficient investment in local data justice infrastructure.
• To achieve climate justice, the author recommends strengthening data sovereignty through local AI infrastructure, democratising governance via public oversight bodies, and fostering South-South cooperation through regional bodies like the African Union and SADC.
• A decolonial approach to climate AI requires 'epistemic pluralism,' which integrates non-technical knowledge sources—such as oral histories and seasonal calendars—as legitimate datasets equal to scientific models.

APA

Bhanye, J. (2025). Artificial Intelligence (AI) and the New Colonialism of Climate Data in the Global South. Public Affairs Research Institute. https://pari.org.za/wp-content/uploads/2025/12/20251205_AI_Johannes-Bhanye_WorkingPaper.pdf

Chicago

Bhanye, Johannes. Artificial Intelligence (AI) and the New Colonialism of Climate Data in the Global South. Public Affairs Research Institute, 2025. https://pari.org.za/wp-content/uploads/2025/12/20251205_AI_Johannes-Bhanye_WorkingPaper.pdf.

Wikipedia

{{cite report |last1=Bhanye |first1=Johannes |title=Artificial Intelligence (AI) and the New Colonialism of Climate Data in the Global South |publisher=Public Affairs Research Institute |date=December 2025 |url=https://pari.org.za/wp-content/uploads/2025/12/20251205_AI_Johannes-Bhanye_WorkingPaper.pdf |access-date=23 October 2026 |via=Climate Insights Directory}}

BibTeX

@techreport{bhanye2025artificial, author = {Bhanye, Johannes}, title = {{Artificial Intelligence (AI) and the New Colonialism of Climate Data in the Global South}}, institution = {Public Affairs Research Institute}, year = {2025}, month = dec, url = {https://pari.org.za/wp-content/uploads/2025/12/20251205_AI_Johannes-Bhanye_WorkingPaper.pdf}, urldate = {2026-10-23}, note = {Indexed by Climate Insights Directory} }