UN report warns AI could widen global inequality, urges nations to build local AI infrastructure

A new United Nations report has raised concerns that artificial intelligence could exacerbate global inequality unless countries adopt a shared framework for its responsible development, as AI adoption and investment remain uneven across the world.

The report highlights that access to AI tools alone does not guarantee equal benefits. Countries that rely on foreign AI models, cloud infrastructure and data pipelines may gain access to the technology while losing practical control over its standards, safeguards and suitability for local needs.

The analysis is part of a report by the Independent International Scientific Panel on AI, established by the UN General Assembly last year as "the first global scientific body on AI." It outlines both the opportunities and risks posed by AI—from its transformative potential in sectors such as agriculture and education to catastrophic outcomes if malicious actors use the technology to commit fraud, spread disinformation or influence elections.

More from Tech

The report also serves as a policy toolkit, offering initial guidance to UN member states on how to harness AI's potential to drive economic growth and innovation across industries while minimizing and addressing associated risks.

Among its recommendations are developing local AI infrastructure, including data centres; improving AI literacy in schools and across the workforce; investing in AI developers and research talent; establishing AI safety institutes; creating strategies to combat AI-driven disinformation; and continuously monitoring how AI systems behave after deployment.

Panel's advisory for countries lagging behind

The panel advises countries that are behind in AI development to make significant investments in computing and data infrastructure. Attracting such investment will require securing a reliable energy supply and expanding data centre capacity. At the same time, the report acknowledges the environmental costs of data centres, including their high energy and water consumption and their contribution to greenhouse gas emissions.

The report also highlights the growing challenge of evaluating the safety of increasingly powerful AI models and ensuring adequate oversight.

"Most countries, including many advanced economies, lack the technical expertise to evaluate cutting-edge 'frontier' AI models or play a meaningful role in shaping their governance," the report states.

Prepared by a panel of 40 independent experts from around the world, the UN-backed assessment describes itself as "the first of its kind," arguing that the UN is "the foremost global forum on transboundary risks of this scale" and that its approach is "scientific, not political."

The report also warns that AI is widening the digital divide through language and connectivity gaps. While generative AI performs well in English and a handful of widely spoken languages, "most languages are either excluded or have much lower performance."

It illustrates the risks with an example of AI mistranslating medical terms in Tigrinya—confusing smallpox with syphilis, gonorrhoea with diabetes, and translating the phrase "you have been given intravenous antibiotics" as "you have been given intravenous insecticides." "These mistranslations can be life-threatening," the report warns.