In its report, Anthropic analyzed how users in more than 150 countries used the Claude language model in November 2025 and confirmed that measures of AI adoption intensity — such as the Anthropic AI Usage Index — remain heavily concentrated in high-income countries. Nations such as the United States, India, Japan, the United Kingdom, and South Korea show significantly higher levels of Claude usage than countries with lower levels of economic development. Unlike within the United States, where regional differences in adoption show some signs of convergence, there is no clear global trend toward faster equalization.
Anthropic emphasizes that this concentration of AI adoption in high-income economies could have tangible economic consequences, as AI tools are increasingly supporting more complex tasks and those requiring advanced technical skills — areas where resources, digital infrastructure, and education are more readily available in developed countries. This means that nations with weaker economic potential and lower levels of skills specialization may fall even further behind in terms of productivity gains and economic growth driven by AI.
Further data from the index indicate that in lower-income countries AI is more often used mainly for education or personal activities, rather than for professional or business applications that could directly contribute to productivity and competitiveness. This may entrench a global divide: while highly developed economies invest in AI to boost efficiency in complex business processes, less affluent countries focus primarily on basic educational uses.
Experts note that without targeted public policies — such as investment in digital infrastructure, education programs, and access to AI tools for businesses — global differences in AI adoption could further exacerbate existing gaps in living standards and economic growth dynamics between countries. The findings of the Anthropic report highlight the need for policy action and public investment to help balance access to and the effects of AI, so that the benefits of the technology are distributed more evenly.