A new United Nations report warns that the development of artificial intelligence may exacerbate global inequality.
As adoption and investment accelerate unevenly worldwide, the report proposes a shared framework for responsible AI development.
“Access to AI tools alone does not produce equal benefit,” the report states.
“Countries that rely on foreign models, cloud infrastructure and data pipelines may gain access to AI while losing practical control over its standards, safeguards and local fit.”
The analysis comes from the independent international scientific panel on AI, established by the UN General Assembly last year as the first global scientific body on AI.
It details both risks and opportunities, from transformative capabilities in agriculture and education to catastrophic outcomes when bad actors deploy AI for fraud or election interference.
Recommendations and Challenges
The preliminary report also serves as a toolkit, offering guidance to UN member states on capitalizing on AI’s potential while minimizing threats.
Suggestions include developing local AI infrastructure, improving AI literacy, investing in developers, building AI safety institutes, combating disinformation, and continuously measuring AI system behavior after release.
The report specifies that measurement should occur “with real users, real tasks and real environments.”
While over a billion people now use AI weekly, access and usage vary widely. Adoption across the global south lags far behind the global north.
The United States and China dominate in developing leading AI models and investing in compute infrastructure.
“The concentration of AI capabilities in a small number of firms and countries could enable authoritarian capture and undermine democratic accountability,” the report warns.
The panel advises lagging countries to invest significantly in computing and data infrastructure, which requires reliable energy and data centers.
However, it acknowledges the environmental costs, including high energy and water consumption and greenhouse gas emissions.
Challenges in evaluating safety and oversight of powerful AI models persist.
“Most countries, including many advanced economies, lack the technical expertise to assess the most capable ‘frontier’ models or to participate meaningfully in their governance,” the authors write.
The panel of 40 independent experts calls this report “the first of its kind.”
They argue the UN is the foremost global forum on transboundary risks, and its approach is “scientific, not political.”
Differences in language and internet access compound the digital divide. “Artificial intelligence leaves most languages behind,” the report notes.
While generative AI performs well in English and other widely used languages, “most languages are either excluded or have much lower performance.”
These disparities can have life-threatening implications, especially in healthcare.
For example, a machine translation of Tigrinya mixed up smallpox with syphilis, gonorrhoea as diabetes, and “you have been given intravenous antibiotics” as “you have been given intravenous insecticides.”
More than 2 billion people—almost a third of the global population—remain completely offline, according to the International Telecommunication Union.
