Islamabad, Nov 11,2026: Artificial intelligence offers significant opportunities to improve productivity across developing economies. Yet, uneven digital infrastructure, workforce skills and technological capabilities could widen economic disparities as countries move at different speeds in adopting the technology.
The World Bank’s World Development Report 2026 estimates that 16.2% of jobs in developing economies could experience meaningful productivity gains from AI, compared with 18.7% in high-income countries.
However, translating this potential into higher economic output will depend on countries’ ability to provide reliable connectivity, computing infrastructure, skilled workers and institutions capable of integrating AI into productive activities.
The World Bank warns that without these complementary capabilities, artificial intelligence could widen productivity gaps between economies rather than narrow them.
The report also estimates that only 4.5% of jobs in low- and middle-income countries face potential automation risks from generative AI, compared with 14.2% in high-income economies.
Prof. Dr. Muhammad Usman Ghani Khan, Director of the National Centre of Artificial Intelligence at UET Lahore, told Wealth Pakistan that the next productivity divide could emerge not from access to AI models themselves, but from differences in countries’ ability to integrate them into economic activity.
“Access to an AI tool is increasingly becoming easier, but productive adoption requires much more. Countries need reliable connectivity, computing infrastructure, skilled people, usable data and organisations capable of redesigning the way work is performed,” he said.
“If those foundations are missing, AI may remain a consumer technology rather than becoming a productivity technology,” he added.
Connectivity data highlight the scale of the challenge facing developing economies.
According to the International Telecommunication Union, approximately 6.3 billion people are using the internet in 2026, leaving around 2 billion people offline. Internet usage has reached 93% of the population in high-income economies but remains at only 32% in low-income countries.
The divide is also pronounced across regions and country groups. Only 42% of Africa’s population uses the internet, while connectivity stands at 41% across least developed countries and 43% in landlocked developing economies.
These differences could influence how quickly businesses and workers benefit from AI-assisted production, services and decision-making, particularly where reliable digital connectivity remains limited.
A joint study by the International Labour Organization and the World Bank covering 135 countries found that developing economies have comparatively low exposure to outright AI automation but substantial potential for worker augmentation.
However, the study cautioned that gaps in digital infrastructure could trigger technological disruption before countries can realise its productivity benefits.
Khan said this imbalance was particularly important for developing economies seeking to use AI to improve economic performance.
“A worker may theoretically benefit from AI assistance, but that benefit cannot materialise consistently if internet access is unreliable, computing is expensive or the workplace has not digitised its processes. Productivity gains require an ecosystem, not simply software,” he said.
The global investment landscape also presents challenges, as capital flows towards AI and digital infrastructure are becoming increasingly concentrated.
UN Trade and Development (UNCTAD) warned in 2026 that investment in AI and digital infrastructure was becoming concentrated in fewer countries and sectors, raising the risk that developing economies unable to attract sufficient capital could fall further behind.
UNCTAD also reported in September that fewer than one-third of developing countries had adopted national AI strategies, while computing capacity, finance, data and technical expertise remained concentrated in a relatively small group of economies and companies.
These trends highlight the importance of building domestic technological capabilities and creating conditions that support the wider adoption of AI across different sectors of the economy.
Khan said developing countries should initially focus on applying artificial intelligence in areas that yield measurable productivity gains without requiring the expensive computing infrastructure needed for frontier models.
“Agriculture, healthcare, education, manufacturing, public administration and financial services all offer opportunities. The objective should be to adapt AI to local languages, datasets and operational problems rather than trying immediately to compete in building the largest models,” he said.
The World Bank similarly argues that developing countries do not need to build the most expensive frontier models to benefit from artificial intelligence. Instead, they need to adapt available technologies to local institutions, languages and development priorities.
Such an approach could help developing economies make practical use of existing AI systems while gradually strengthening their digital infrastructure, technical expertise and institutional capacity.
Khan said the central policy challenge was increasingly to ensure that AI capabilities reached a broader range of businesses, workers and public institutions rather than remaining concentrated in technologically advanced sectors.
“If AI capability becomes concentrated only in leading firms, major cities or technologically advanced countries, productivity gaps can widen very quickly. The real development objective should be to spread useful AI capacity across businesses, workers and public institutions,” he said.
For developing economies, the global AI transition therefore presents both an opportunity to improve productivity and a challenge to ensure that technological progress does not deepen existing economic disparities.
The long-term benefits will depend less on which countries gain access to AI tools first and more on their ability to develop the infrastructure, skills, data and institutions needed to turn technological adoption into sustained productivity growth.