Artificial intelligence could entrench a new era of global inequality unless governments act to make its benefits widely accessible, a new United Nations report has warned.
The study by the UN Development Programme (UNDP) says most of the gains from AI are likely to accrue to wealthier countries, mirroring the “Great Divergence” of the industrial revolution when Western economies surged ahead and others fell behind.
Launched in Bangkok on Tuesday, the report — The Next Great Divergence: Why AI May Widen Inequality Between Countries — argues that AI is a “general-purpose technology” capable of lifting productivity, creating new industries and helping latecomer economies catch up. But it stresses that such outcomes are not guaranteed, and depend on basic preconditions such as reliable electricity, internet access, skills and effective institutions.
The authors point out that about a quarter of people in the Asia–Pacific region still lack online access, leaving millions without the connectivity needed to benefit from AI-powered services, from telemedicine to digital banking and online education. Within countries, they say, regional and social inequalities mean that even advanced economies contain communities at risk of being left further behind.
Michael Muthukrishna of the London School of Economics, the report’s lead author, said debates around AI often focus on headline-grabbing advances while neglecting their impact on everyday life.
“We tend to overemphasise the role of technology,” Muthukrishna told the launch event via video link. “We need to ensure it’s not technology first, but people first.”
The report highlights stark contrasts in preparedness. Countries such as China, Japan, South Korea and Singapore are described as well placed to deploy AI at scale, thanks to strong digital infrastructure, skills and investment. By contrast, states including Afghanistan, the Maldives and Myanmar lack the power systems, connectivity and human capital needed to harness advanced computing, raising the risk that they will see few direct benefits from the current wave of innovation.
UNDP also links the AI debate to wider development trends. Its 2025 Human Development Report, A matter of choice: People and possibilities in the age of AI, found that global progress in health, education and income has slowed to its weakest pace in 35 years, even as AI adoption accelerates. The agency warns that without deliberate policy choices, earlier gains in closing the gap between richer and poorer nations could be eroded.
Economically, the potential is substantial. One analysis cited by UNDP suggests AI could add around two percentage points to annual GDP growth in parts of Asia, with ASEAN economies alone gaining close to $1 trillion in additional output over the next decade. Yet the same modelling indicates that jobs held by younger workers and women may face higher exposure to automation if education, retraining and social protections do not keep pace.
The report lists a series of risks that could fall disproportionately on vulnerable groups. Populations displaced by war, civil conflict or climate-related disasters, as well as older people and marginalised communities, may be missing from the datasets used to train AI systems, making them effectively “invisible” to algorithms that guide decisions on welfare, credit or healthcare. In addition, opaque “black box” systems can encode and amplify existing biases, particularly where oversight is weak.
Environmental impacts are another concern. Large data centres require substantial electricity and water, raising questions in both rich and poor countries over their contribution to carbon emissions and pressure on local resources. Expanding power generation to serve AI-driven demand may complicate efforts to meet climate targets if it relies on fossil fuels.
UNDP flags mounting ethical, privacy and security issues. Researchers have already documented the use of AI tools to automate parts of cyberattacks, while generative systems have made it easier to produce convincing “deepfake” audio and video that can be used to mislead voters, blackmail individuals or facilitate fraud. The report also points to the expansion of AI-enabled surveillance, which can infringe privacy and civil liberties when deployed without clear legal safeguards.
Philip Schellekens, UNDP’s chief economist for Asia–Pacific, said governments should treat AI as basic infrastructure for modern life, comparable to electricity and the internet. He called for increased investment in digital networks, education and training, as well as competition policy and social safety nets, to ensure that AI-driven disruption does not fall unevenly on those least able to absorb it. “We believe we need more balance, less hysteria and hype,” he said.
The report concludes that the direction of travel is still a matter of political choice. It urges states to adopt transparent governance frameworks, strengthen data protection and build international co-operation so that poorer countries are not locked out of the emerging AI economy. “The goal,” it says, “is to democratise access to AI so that every country and community can benefit while protecting those most at risk from disruption.”
