The United Nations has issued a stark warning that the rapid, uneven spread of artificial intelligence technologies threatens to entrench and even widen global inequality. In a report released Monday, the UN's technology advisory body found that 87% of AI research papers, 94% of AI patents, and virtually all frontier model training compute remains concentrated in North America, Europe, and China.
The report highlights a cascade of compounding disadvantages: developing nations lack the compute infrastructure, the training data in local languages, the regulatory expertise, and the talent pipelines to participate meaningfully in the AI economy. Meanwhile, AI-driven automation threatens to eliminate precisely the low-cost-labor advantages that have historically enabled developing economies to grow.
'We are at risk of creating a world where a handful of companies and countries capture the entire value of the AI revolution while the rest of humanity is left to cope with its disruptions,' said the report's lead author. The UN is calling for a global AI fund — financed by a small levy on AI company revenues — to invest in compute infrastructure, education, and local-language data initiatives across the Global South.
The report also singles out the concentration of AI compute as a national security concern, noting that 12 of the world's 15 largest GPU clusters are located in a single country. It recommends that international development banks prioritize AI infrastructure lending and that frontier model developers be required to reserve a percentage of training compute for publicly beneficial research in underrepresented regions.
Source context: United Nations
