Read Also

Job vacancies requiring AI skills grew fourfold in Kenya between 2021 and 2024, a World Bank report now indicates, as digital skills continue to form the baseline across many occupations.

As developing countries continue to navigate AI adoption, 40 per cent of ChatGPT’s global traffic came from middle-income countries in mid-2025.

Conversely, Kenya, Nigeria and the Philippines saw a tripling of vacancies by mid-2023, maintaining levels above those of early 2021, with strong trends continuing into 2025 and 2026, albeit with recent moderate declines.

“Vacancies doubled in the Arab Republic of Egypt, Pakistan, and the Philippines, and grew fourfold in Kenya, albeit from a very low base. Job vacancies requiring GenAI skills surged ninefold globally from 2022 to 2024,” the World Bank Digital Progress and Trends Report reads in part.

The IT industry is most likely to require AI skills across countries, followed by professional services and finance.

“The IT industry demonstrates the highest probability of AI skills demand across most countries, with around eight per cent of IT industry job postings requiring AI skills globally,” the World Bank notes.

Professional services and finance are the other top industries that require AI skills, with about four to five per cent of job postings requiring such competencies.

Managers and professionals in manufacturing are also increasingly expected to possess AI skills.

These skills are, however, less required in the agriculture industry.

In occupations, software developers, database designers and administrators, mathematicians, and physicists are most likely to require AI skills, with more than 10 per cent of vacancies requiring them.

Since 2022, Nigeria, Morocco, South Africa, Egypt, and Kenya have recorded the highest cross-border migration of high-skilled talent in Africa.

Digital skills are now a baseline across many occupations, with AI reshaping labour markets globally.

Skills gaps are uneven, but AI-related jobs are growing faster in middle-income countries than in high-income countries.

“Many low-income countries and middle-income countries face significant talent challenges from supply shortages—due to the lack of quality, industry-aligned education and training programs—to brain drain, hindering full participation in the AI economy,” World Bank notes.

Brain drain further challenges AI development and adaptation, with talent outflows three to four times higher than inflows in countries like Bangladesh, Lebanon, Nigeria, and Ukraine.

According to the World Bank, adapting global models to local economic, cultural, and institutional contexts can turn global data resources into locally relevant AI solutions.

About 85 per cent of AI training data start-ups are based in high-income countries—leaving low and middle-income countries behind in localized data.

Venture capital investment in AI training data is led by the United States at 56 per cent, followed by China at 17 per cent, the European Union at 15 per cent, India at three per cent, and the United Kingdom at two per cent, with the rest of the world accounting for the remaining six per cent.

More than 50 per cent of open-source AI training data sets are in English, limiting accessibility for non-English speakers.

Emerging formats like video and audio offer new ways for LICs and MICs to participate. For example, eight per cent of YouTube videos are in Hindi, and three per cent are in Arabic.

Meanwhile, LICs such as Ethiopia and Kenya underperform on both scale and intensity, according to the World Bank, highlighting the need to invest in foundational enablers such as affordable and high-quality broadband, digital literacy and workforce skills.

“For these countries, building readiness means prioritising the basics that unlock broader AI adoption.”