Africa’s AI Policy Ambitions Ignore Energy, Climate and Labor Concerns

AI strategies across Africa are becoming increasingly ambitious in response to the continent’s growing drive for AI development. This ambition is evident in plans to establish new data centers and expand computing capacity. Currently, at least 226 data centers are operational across 39 African countries, with many strategies outlining the goal of building more.

Examples include Benin’s intention to upgrade its data center to meet AI compliance standards and Egypt’s plan to construct a “cutting-edge domestic data center.” While this momentum toward expanding AI capacity is understandable, it reveals a significant limitation: the near neglect of issues related to energy use, environmental impact, and labor exploitation.

This finding is based on an August 2025 analysis of 14 publicly available AI strategies, both finalized and draft editions, released by the African Union, Benin, Egypt, Ethiopia, Ghana, Kenya, Lesotho, Mauritania, Mauritius, Nigeria, Rwanda, Senegal, South Africa, and Zambia.

African governments must address this oversight, not least because of the message it sends to AI developers both within and beyond the continent, namely, that concerns about energy, climate, and labor can be sacrificed on the altar of unchecked AI advancement.

The costs of AI development

AI development comes with certain costs, particularly in relation to labor, energy, and the environment, and these have been widely documented. It is known, for instance, that AI systems and data centers threaten the world’s fresh water supply and are estimated to consume as much energy by 2027 as a country like the Netherlands would require.

This represents a perilous demand for valuable resources, including the mining of rare earth metals, with consequences for the environment. Added to these are labor exploitation concerns, especially in global majority countries like Kenya, where data workers are underpaid and operate in unsafe mental health conditions.

These concerns should ordinarily worry African governments, given that the continent suffers from key infrastructural deficits. For instance, 85% of people worldwide who lack access to electricity live in sub-Saharan Africa and the continent is disproportionately affected by climate change costs and water shortages. The strategies, however, show a likely disregard for these concerns.

Limited attention to energy and climate concerns

My analysis of the 14 AI strategies reveals that most countries in Africa intend to increase their data extraction and processing for AI development. Yet only six strategies (Egypt, Ethiopia, Kenya, Mauritania, Mauritius, and Nigeria) mention energy concerns or the need to provide alternative energy sources.

Some of the alternatives mentioned include energy efficiency measures, notable in the Egyptian and Mauritian strategies, and the use of green and renewable energy sources, evident in the Nigerian and Mauritanian strategies. Ethiopia, on its part, sees the need to increase its energy generation capacity by strengthening its hydroelectric potential.

Even fewer strategies (three out of 14) acknowledge the environmental consequences of increased AI advancement. This can be seen in the AU Strategy, which notes that “the high demand for fresh water to cool data centers poses a threat to regions already facing water scarcity.”

Kenya also states that the energy requirement of LLMs and data centers has “long-term environmental impacts on Kenya’s natural resources.” And although they highlight the concerns, nothing is said about how to address them, except in the Senegalese strategy, where policymakers articulate the need to “integrate the question of the environmental impact of AI into the environmental code.”

Other than these three cases, the few strategies that mention AI in its environmental contexts only specify the applied uses or benefits of AI in the mining sector (Ethiopia), wildlife conservation (Ghana), and climate change (Mauritius).

Deprioritizing AI labor exploitation

When it comes to labor, it is well known that AI downstream workers, such as data labelers and annotators, who work under exploitative conditions, predominantly reside in Africa and other global majority regions. Most of the strategies, however, only highlight job displacements tied to automation, outlining the need to address them through retraining and upskilling. Examples include the strategies of Egypt, Lesotho, Mauritius, Nigeria, and Zambia.

Countries like Ghana and Rwanda actually encourage the practice of data labeling in their countries, without acknowledging the exploitation that African labellers face. Ghana, for instance, sees the need to “future-proof” the workforce, adding that the “youth can participate in data collection, labelling, applied data science, machine learning research and beyond.” Rwanda also hopes to establish local data value-chains, which will begin with activities such as data annotation and labeling.

Kenya is the only country that directly mentions the exploitation that labelers face, noting that “Many Kenyans work in AI but remain stuck in bottom-of-pyramid and entry-level jobs such as data annotation.” It specifically notes that “some of these workers are already supporting international companies such as Sama in the development of AI through outsourced data processing and labeling.”

By naming Sama, which is embroiled in labor exploitation controversies, Kenya shows that it recognizes the weight of the issues at stake. Yet, the strategy only tacitly highlights the need for better career progression, saying nothing about how to address exploitative AI labor.

Pursuing AI expansion without neglecting key concerns

While it is clear that AI strategies are not legally binding, they nonetheless convey the government’s priority agenda for the development, deployment, and governance of AI systems. Hence, the negligible attention to concerns around energy and the environment potentially sends a signal to AI developers within and outside the continent that these concerns can be ignored in the rush for rapid AI advancement.

One can understand why African governments are eagerly pursuing AI development in relation to talent, data, and infrastructure: they face the reality of losing out in the so-called race for AI development. The drive towards data sovereignty also warrants that African countries and corporations invest in building infrastructures such as data centers locally. But these should not be pursued without significant consideration of the adverse impact that they have on energy supplies and the environment.

When it comes to labor exploitation, the impact on data workers is clear, and the fact that this is not addressed in the strategies sends a clear message to AI companies: African governments are willing to overlook the concerns so long as investments come in. It reflects the desperation tied to high unemployment in many African countries, and the openness to job creation, even if the opportunities are exploitative and reflect historical patterns of colonial extraction.

This situation is not helped by the realization that a good number of AI strategies across Africa are usually developed in close collaboration with external stakeholders. These typically include the German International Development Agency (GIZ), the UK’s Foreign, Commonwealth and Development Office, the European Union, and the now-defunct US Aid for International Development – one or more of which were involved in developing the AI strategies of Ghana, Kenya, Lesotho, Rwanda, Senegal, and Zambia. While there is some value in drawing from external expertise, it is vital that African policymakers are not overly swayed by it and instead focus on realities and actual concerns that relate to the African experience.

The silver lining is that the concerns around data centers, energy, labor, and the environment can be addressed in future editions of the 14 strategies. African countries that have not developed their strategies are even more uniquely placed to address these concerns and send the right signals to AI developers operating within and outside the continent.