UN Panel Warns AI Development Could Deepen Global Inequality

UN Scientific Panel Flags AI Inequality Risk

A newly released United Nations report cautions that artificial intelligence development threatens to widen existing global inequalities as adoption and investment accelerate unevenly worldwide. The analysis comes from an independent international scientific panel established by the UN General Assembly in 2025, which describes itself as the first global scientific body focused on AI.

The report emphasizes that simply having access to AI tools does not guarantee equal benefits. Countries dependent on foreign models, cloud infrastructure, and data pipelines may use AI without retaining meaningful control over standards, safeguards, or local customization.

Why It Matters

As AI becomes embedded in critical sectors from healthcare to agriculture, the concentration of capabilities in a handful of firms and nations could enable authoritarian control and weaken democratic oversight. For business leaders, this signals that AI strategy must account for infrastructure sovereignty and local adaptation—not just tool access.

Concentration of Power and Resources

The United States and China currently dominate both the development of leading AI models and investment in compute infrastructure—the hardware, memory, networking, and storage necessary to run advanced systems. This concentration creates structural disadvantages for other regions.

While more than one billion people now use AI weekly, adoption across the global south significantly lags behind the global north, according to the panel's findings. The report notes that most countries, including many advanced economies, lack the technical expertise to evaluate the most capable frontier models or participate meaningfully in their governance.

Language and Access Barriers

Language disparities compound the digital divide. Generative AI tools perform well in English and other widely used languages but exclude or underperform in most languages. The report highlights a dangerous example: machine translation of Tigrinya confused smallpox with syphilis, gonorrhea with diabetes, and translated "intravenous antibiotics" as "intravenous insecticides"—errors that could prove life-threatening in healthcare settings.

Internet access remains another fundamental barrier. Nearly 2 billion people—almost one-third of the global population—remain completely offline, according to the International Telecommunication Union.

Framework for Responsible Development

The panel's preliminary report functions as both analysis and toolkit, offering guidance to UN member states on capitalizing on AI's potential while addressing risks. Recommendations include developing local AI infrastructure such as data centers, improving AI literacy in education and workforce training, investing in developer communities, establishing AI safety institutes, creating disinformation countermeasures, and continuously measuring how AI systems behave with real users in real environments.

The report acknowledges the environmental costs of data centers, including substantial energy and water consumption and potential greenhouse gas emissions. For countries seeking to build AI capacity, the panel advises significant investment in computing and data infrastructure, which requires securing reliable energy supplies.

The 40-member panel of independent scientific experts from around the world characterizes this as the first report of its kind, positioning the UN as the primary global forum for addressing transboundary risks of this scale through a scientific rather than political approach.

These details were first reported by The Guardian.