Will AI help or hurt Africa?

At the annual World Bank-International Monetary Fund Annual Meetings in Washington, DC, this week, delegates will discuss how AI could “unlock” opportunities in developing nations. This was also a hot topic at last month’s UN Summit of the Future. And nowhere is that discussion more ripe than the African continent.

AI technology might be able to help poorer nations “leapfrog” entire development phases, the Financial Times wrote this week — just like how some nations skipped mass landline adoption and went straight to mobile phones in the last two decades.

AI startups are popping up across Africa, trying to tackle problems in health care, education, and language and dialect differences. And foreign firms are starting to invest too: Microsoft and the UAE-based fund G42 announced a $1 billion investment in Kenya to build data centers, develop local-language AI models, and offer skills training to people in the country. Amazon has said it’s investing $1.7 billion in Amazon Web Services cloud infrastructure across the continent. For its part, Google has begun developing African-language AI models and given $6,000 microgrants to Nigerian AI startups.

But there’s also concern that AI could deepen existing digital divides — especially if popular large language models aren’t developed with Africa in mind, don’t support local languages, or if the continent lacks the infrastructure to run high-powered models efficiently.