Artificial intelligence is projected to generate $2.9 trillion for Africa by 2030, along with half a million new jobs each year and a path out of poverty for more than 11 million people, according to Yves Iradukunda, permanent secretary at Rwanda’s Ministry of ICT and Innovation.

That bold vision set the tone at Africa’s first global AI summit, held April 3–4 in Kigali, where researchers, scientists, and policymakers from over 90 countries gathered to shape the continent’s digital future.

“Africa must take a selective, strategic approach tailored to its real capabilities,” said Togolese President Faure Gnassingbé, the summit’s guest of honor. He emphasized health care, education, and agriculture as the three pillars of Africa’s AI development strategy.

Closing the data gap

But realizing that vision will require major infrastructure improvements. Many of the continent’s digital datasets are stored outside Africa, limiting access and sovereignty over its own information.

“We face serious problems not only accessing data but collecting our own,” said Ikram Abdirahman, a Kenyan lawyer specializing in AI policy.

A late 2023 International Monetary Fund report ranked sub-Saharan Africa as the world’s least-prepared region for AI development. Only five countries — South Africa, Nigeria, Kenya, Egypt, and Morocco — currently host AI-focused data centers.

Fears of a digital colonization

Beyond infrastructure, leaders warn that Africa risks becoming a passive consumer of AI systems shaped by Western data and biases. Abdirahman said using foreign training data could reinforce harmful stereotypes, including in health care, where many AI applications fail to address diseases prevalent on the continent such as malaria, Ebola, and the Marburg virus.

In November 2023, Seydina Ndiaye, a member of the UN AI think tank, warned of a “new colonization of Africa” through AI. A 2024 UN report echoed that concern, stating that AI systems could contribute to racial discrimination if not properly localized.

Linguistic and cultural inclusion

Localizing AI is also a cultural imperative with more than 2,000 languages spoken across Africa. Francis Kombe, CEO of EthiXPERT (the NGO focusing on enhancing research integrity, technology, and innovation, particularly in the field of artificial intelligence) and chair of the African Research Integrity Network, called for building a dedicated linguistic infrastructure for African AI development.

“The large language models available today were created for the Global North,” Kombe said. He advocates training AI systems in local languages — starting with Kiswahili — and gradually expanding to community languages.

That vision aligns with a June 2024 declaration by 130 African tech ministers, who called for AI systems that reflect Africa’s linguistic, cultural, and historical diversity.