The rapid spread of artificial intelligence risks leaving millions of people further behind unless countries have the skills, systems and safeguards needed to shape how the technology is used, the United Nations Development Programme (UNDP) has warned.
The agency said the divide is especially clear in communities whose languages and local realities are poorly represented in AI systems, raising concerns that access to the technology alone may not be enough to ensure people benefit from its growing use.
In an update posted on X on August 18, 2026, the UNDP said generative AI was being used by 17.8 percent of the world's working-age population by the first quarter of 2026, up from 16.3 percent in the second half of 2025.
However, adoption remains much higher in richer parts of the world, with usage standing at 27.5 percent in the Global North compared with 15.4 percent in the Global South.
The agency said the growing gap reflects more than differences in access to technology. It also affects who has the ability to influence how AI is developed, introduced and used in economies, public institutions and communities.
Twenty-six economies have now recorded generative AI usage of more than 30 percent among working-age adults.
The UNDP also cited World Bank estimates showing that high-income countries have about 50 times more ChatGPT users than low-income countries, while fewer than one in five workers in low-income countries are exposed to AI.
The agency said language is among the clearest signs of the unequal development of AI. Of more than 7,000 languages spoken around the world, fewer than 100 are effectively supported by AI systems.
“Most African languages, along with many in South and Southeast Asia, the Pacific and Indigenous communities, remain underrepresented,” the UNDP said.
It warned that communities with limited representation in AI systems could struggle to benefit from AI-powered services and new economic opportunities as the technology becomes more widely used.
“Which means people may have AI tools without receiving the benefit. Systems may misunderstand needs, overlook local knowledge or leave communities out of next generation digital services and economic opportunity,” the UNDP said.
The agency said the key issue is therefore no longer only how quickly AI is spreading, but also who controls and influences its use and who gains from its adoption.
“The question is no longer simply how fast AI spreads, but who influences how it is used and who benefits from its adoption,”
According to the UNDP, decisions on data, infrastructure, governance, skills, procurement and public oversight will play a major role in determining whether AI creates more opportunities and gives people greater control over their lives or adds to existing inequalities.
The agency said countries do not have to become fully self-sufficient in technology to have a say in the future of AI. Instead, they need enough local capacity to understand the technology, assess its risks and make informed decisions within an increasingly connected global AI system.
The UNDP is supporting countries through AI landscape assessments, national AI strategies, trust and safety programmes, local-language ecosystems and efforts to strengthen capacity in the public sector.
Its AI Sprint initiative currently supports more than 60 interventions in over 40 countries, while its Strategic Plan for 2026-2029 lists digital and AI transformation among the key areas that can speed up development.
The agency also called for safeguards as AI becomes more common in public services, financial systems and government administration.
“Countries need practical ways to ask better questions before systems are adopted, monitor how tools perform, respond when something goes wrong, and make sure people can challenge or correct outcomes,” the agency said.
The UNDP said the direction AI takes has not yet been decided, but choices being made now will shape how the technology affects people's lives in the years ahead.
It said human development will depend on whether countries build the institutions, infrastructure, skills and safeguards needed to make sure AI serves people instead of widening inequalities that already exist.