World Bank Urges Developing Nations to Invest in Local AI
Global leaders and technology experts said such tools could create jobs and improve public services. They also expect these tools to widen economic opportunities and help people start businesses.
Work and Business Shift
Senior Singaporean leader Tharman Shanmugaratnam said easier access to AI is changing how people work. He believes the gap between employees and entrepreneurs is starting to fade.
World Bank President Ajay Banga and IMF chief Kristalina Georgieva joined the discussions. The talks centred on "Small AI", meaning cheap tools built for specific tasks.
What Small AI Offers
Small AI tools can run on basic devices and in areas with weak internet connections. Participants said they could support farming, healthcare, education, skills training, and small businesses.
Speakers urged countries to move beyond pilot projects and use the technology in daily life. They called for investment in digital infrastructure, training, local data, and responsible governance.
Africa's Language Gap
The World Bank's Africa Economic Update, published in October 2026, raises concerns about African representation in AI training data. Africa has about 2,000 languages, yet roughly 56 per cent of open-source training data is only in English.
This imbalance may limit how well AI systems understand African users. The report says the deepest returns come from adapting AI to local data, languages, and problems.
A Local Success Story
The report highlights InkubaLM, a small language model built by African company Lelapa AI. Developers trained it on isiZulu, Yoruba, Hausa, Swahili, and isiXhosa.
The model reportedly beats much larger systems in those languages. This shows that tools designed for specific communities can outperform bigger global platforms.
Barriers to Progress
Adding African languages to existing platforms is not enough, the report warns. Weak digital infrastructure, limited technical skills, and poor access to computing power still block progress.
The World Bank also wants clear rules on data ownership, privacy, consent, and benefit-sharing. Communities that contribute language data should share in the benefits.
Measuring Success
Regional cooperation could help countries share expertise and cut the cost of building tools. This matters most for languages with smaller digital markets.
The report says success should depend on how accurately and affordably people can use AI. It rejects counting the number of languages a system claims to support.
Post a Comment