AI in Developing Countries: Opportunities and Risks
Artificial Intelligence (AI) holds the potential to trigger a "new era of divergence" where the gaps between countries widen rather than shrink. This is according to a recent report by the UN Development Programme (UNDP) highlights. The question of who benefits from this technology and who is left behind is becoming increasingly political. The following sections shed light on the power imbalance in the distribution of data centers, data, and know-how, the impact of AI on developing countries, and the importance of fair access to AI.
Source: Own Representation
AI & Global Inequality
International organizations have changed their tune regarding AI. The UNDP warns that unchecked AI can deepen economic, social, and political disparities between countries, as the starting point for infrastructure, education, and regulation is extremely unequal. The Human Development Report 2025 from the UNDP focuses on how AI can accelerate development or entrench inequality, depending on access to computing power, data, and skills. Already, the Human Development Report 2023/2024 noted that progress in health, education, and income after the pandemic has been uneven, and the gap between rich and poor countries is widening again.
The new AI Diffusion Report from Microsoft shows that AI is spreading faster than previous foundational technologies, while billions of people remain excluded due to a lack of networks, electricity, and digital skills ( Business Insider). ). The WTO estimates that AI could increase global trade by about a third by 2040, but warns that without inclusive access, the global prosperity gap is likely to widen ( Reuters). ). Analyses by UNDP, WTO, and other institutions show that the lion's share of economic gains will remain in highly developed economies if the distribution of AI infrastructure and skills does not change.
The UNDP report "The Next Great Divergence: Why AI May Widen Inequality Between Countries" argues that AI can reverse the convergence that many countries have benefited from through trade and technology. The report highlights three points:
- Unequal Starting Conditions: Countries like Singapore, South Korea, or China are investing heavily in data centers, fiber optic networks, and AI competencies, while other nations struggle to ensure stable internet and electricity supply ( UNDP).
- Unbalanced Value Creation: A large part of AI innovation is concentrated in a few highly developed economies, a fact that the World Bank also confirms..
- Regulatory Gaps: Most countries lack robust governance frameworks for AI, which exacerbates risks such as discrimination and unemployment ( hdr.undp.org).
Although the UN Security Council and the UN General Assembly have adopted resolutions for safe, human-rights-compliant, and inclusively accessible AI, these are not legally binding. The International Monetary Fund (IMF) warns that many countries are neither regulatorily nor ethically prepared for the AI boom, and the gap in AI capabilities is rapidly growing.
Source: YouTube
Impacts on Developing Countries
The impacts of AI on developing countries can be divided into the areas of labor, statehood, and daily life.
Labor
A growing part of the AI industry relies on poorly paid "invisible" labor in the Global South. Research shows that workers in Kenya were used to train and clean data for large language models, earning less than US$2 per hour for moderating highly distressing content ( TIME, Business Insider). ). Affected individuals reported traumatic experiences from daily exposure to depictions of violence and abuse ( The Guardian, TIME). ). Think tanks such as the Center for Global Development point out that poorer countries often lack the means for retraining programs or social safety nets to cope with job losses due to automation. Without countermeasures, there is a risk that a part of the population will be used as cheap data labor in the short term, while in the long term, many tasks will be automated without new, better-paying jobs emerging.
Statehood and Public Services
AI can make administrations more efficient, but requires a minimal base of digital infrastructure and skilled personnel. UNDP analyses show that many countries in the Global South fail at this threshold due to a lack of data platforms, computing capacity, and regulated access to cloud services ( hdr.undp.org).
Daily Life and Participation
AI also opens up new opportunities for resource-scarce contexts, for example, in medical diagnostics or translation services ( World Bank). ). However, the World Bank shows that many systems are trained on a few languages and data from wealthy countries, meaning people in rural regions of Africa, Asia, or Latin America are often not represented or are poorly represented.
Source: haufe-akademie.de
Source: YouTube
AI Infrastructure & Power Distribution
The debate surrounding AI infrastructure, data centers, and global power distribution determines who can use or further develop modern AI. Powerful AI models require enormous computational resources, which are concentrated in a few countries. Microsoft's AI Diffusion Report shows that countries like the USA, China, Singapore, or the United Arab Emirates are pioneers in AI use and infrastructure, while many countries in Sub-Saharan Africa, parts of South Asia, and Latin America remain significantly below a double-digit usage rate ( Business Insider).
An example is the distribution of data centers: According to a Reuters analysis , the African continent has less than 1% of the world's data center capacity. The World Bank is responding with investments in companies like Raxio, which is building data centers in countries like Uganda, Angola, or Ethiopia ( Reuters). ). Massive AI infrastructure projects are emerging in Asia, such as China's $37 billion data center cluster ( Tom's Hardware). ). In Malaysia, data centers are growing in Johor, primarily serving foreign tech corporations and creating few highly skilled jobs, which critics describe as "digital coloniality" ( AP News).
The World Bank summarizes the dilemma for many developing countries: they must decide whether to build their own data centers or rely on outsourced cloud infrastructure, with value creation and control over data potentially remaining with foreign providers in both cases ( World Bank). ). The WTO warns that AI could not only change trade flows but also concentrate control over critical hardware in a few countries ( Financial Times).
Fair Access & Solutions
Fair access to AI in poor countries requires concrete measures. Various institutions propose a roadmap for more inclusive AI development:
- Infrastructure as a Global Development Task:: The World Bank states that digital infrastructure has become a central development task ( World Bank). ). This includes new financing instruments that allow countries with limited budgets to build or purchase computing capacity.
- Open and "Small" AI Instead of Just Mega-Models:: UNDP and the World Bank emphasize that developing countries can benefit more in the short term from smaller, specialized AI solutions that require less computing power ( hdr.undp.org, Open Knowledge Repository). ). Open-source models and locally adapted systems can reduce dependencies.
- Building Skills and Regulation Together:: UNIDO argues that countries should build their own innovation and production capacities in AI. Without their own experts, nations remain in a weaker negotiating position. The IMF urges that many countries are lagging in regulation and cannot assess the consequences of AI systems.
- Global Rules and Real Participation:: The CSIS analysis emphasizes that the current decisions on infrastructure, governance, and localization will determine whether AI supports or hinders development in the Global South. This requires that lower-income countries be perceived as equal partners. Corporate Responsibility:
- The private sector plays a central role.: Analyses by Brookings show how heavily global AI platforms rely on data work in low-wage countries and how little protection these workers have. Companies must be measured by whether they pay fair wages, provide psychological support, and offer opportunities for advancement. An infographic shows three line charts on the use and planning of AI in companies. An infographic shows three line charts on the use and planning of AI in companies.
Source: user-added
new UNDP report
, the warnings from the WTO, , World Bank, , and IMF, as well as studies on working conditions in the Global South, paint a consistent picture: if rich countries and large tech corporations set the rules alone, many nations will become mere data providers and markets for AI products. At the same time, this is not inevitable. AI can accelerate development processes, improve access to education and healthcare, and foster new industries ( hdr.undp.org
, World Bank, ). Whether this leads to a new wave of participation or a "new era of divergence" depends on political decisions, international rules, and concrete investments in infrastructure, skills, and fair working conditions. The central question is not whether AI is coming, but for whom it is being built and deployed.).