I’ve been tracking the artificial intelligence for development (AI4D) conversation for years, and I have to call out what’s become an increasingly troubling pattern: we’ve reduced the entire Global South to just one continent.
A recent comprehensive analysis by researchers Sophie Toupin and Roda Siad confirms what I’ve observed firsthand—despite AI4D supposedly addressing challenges across the Global South, the vast majority of publications and initiatives focus exclusively on Africa.
This geographic tunnel vision isn’t just intellectually lazy. It’s a massive strategic error that ignores billions of people who could benefit from thoughtful AI4D approaches.
Sign Up Now for more ICT4D commentary
AI4D Only Sees Africans
Let me put this in perspective with some basic arithmetic.
Africa has approximately 1.4 billion people. That’s significant, but it’s not the whole story. India alone has 1.64 billion people. Add Indonesia (288 million), Pakistan (259 million), Bangladesh (178 million), Brazil (214 million), and the Philippines (118 million), and you’re looking at over 2.7 billion people who live in Global South countries outside Africa.
Yet the Toupin and Siad research reveals that the focus remains purely continental rather than truly global.
This pattern extends across academic literature, policy frameworks, and funding mechanisms. They found that while AI4D ostensibly applies “broadly to the global South,” most work uses the term “in reference to AI development on the African continent.” This narrow framing has created what I call geographic myopia in our field.
Asia is Leading AI Adoption
The geographic bias becomes even more problematic when you examine the sophisticated AI strategies emerging across Asia and Latin America.
According to recent global AI deployment statistics, India leads the world at 59% company-level AI deployment, followed by the UAE at 58% and Singapore at 53%. Latin America as a region sits at 47% deployment rates. Compare this to the United States at just 33%.
India recorded the highest rise in AI hiring in 2024 with a 33.4% increase, followed by Brazil and Saudi Arabia. More than 40% of ChatGPT’s global traffic now originates in middle-income countries, led by Brazil, India, Indonesia, and Vietnam. These aren’t countries waiting for AI4D salvation—they’re already at the forefront of AI adoption and innovation.
- India has committed $1.25 billion to expand AI capabilities and established comprehensive national AI frameworks.
- Brazil has launched AI-enabled health diagnostics systems and precision agriculture initiatives.
- The Philippines has built a thriving business process outsourcing sector that’s now adapting to AI integration.
- Bangladesh is pioneering AI integration in garment manufacturing, though this raises important questions about workforce displacement.
These aren’t scattered pilot projects. They represent systematic, government-led approaches to AI integration that rival or exceed initiatives we celebrate in African contexts.
The Cost of Geographic Tunnel Vision
This Africa-centric framing carries three major costs for the AI4D community:
1. We’re missing massive learning opportunities.
When Brookings researchers examined AI adoption across the Global South, they found diverse use cases spanning banana disease identification for farmers, cassava disease diagnosis in East Africa, precision agriculture in Brazil, telehealth engagement in rural India, and clinical decision support in Ghana. The solutions emerging from India and Brazil are just as innovative as those from Kenya or Nigeria, but they receive fraction of the AI4D attention.
2. We’re perpetuating a colonial framing.
Is Africa uniquely dependent on external AI4D assistance while Asian and Latin American innovation is somehow separate from development concerns? This reinforces exactly the type of power imbalances that critical AI4D scholars warn against.
3. We’re limiting the impact of our funding and programming.
When major donors design AI4D initiatives with primarily African focus, they’re making strategic resource allocation decisions that affect billions of people. The recent AI4D Africa program funded by Canada exemplifies this issue. Significant resources dedicated to one continent while ignoring equally pressing challenges across Asia and Latin America.
Beyond Continental Categories
I’m not arguing for less attention to Africa’s AI development needs.
I’m arguing for recognition that AI4D challenges are global, not continental. The five frameworks that Toupin and Siad identify apply across all Global South contexts, not just African ones.
- Developmentalist,
- Economic development,
- International policy,
- Colonial/extractivist,
- Decolonial
The decolonial AI discourse, for example, draws heavily from Latin American scholarship and Indigenous approaches. The economic development framework is perhaps most advanced in India and Brazil. The colonial/extractivist critique applies to how major tech platforms extract data and labor from countries across Asia, Africa, and Latin America.
Recommendations for True AI4D
Moving forward, our AI4D frameworks and funding should reflect that the Global South includes 6.8 billion people across multiple continents. Until we move beyond geographic myopia, we’re not serving the full scope of global development challenges that AI might help address.
The future of AI4D must be truly global, not just African.
The billions of people living in Asia, Latin America, and other Global South regions deserve equal attention in our research, funding, and policy frameworks. It’s time to expand our vision to match the scope of our stated mission.
