Anthropic Study Tracks AI Adoption Across Countries, Industries
- By John K. Waters
- 09/17/25
Adoption of AI tools is growing quickly but remains uneven across countries and industries, with higher-income economies using them far more per person and companies favoring automated deployments over collaborative ones, according to a recent study released by Anthropic.
The report analyzes 1 million Claude.ai conversations in August and a large sample of enterprise API traffic. It introduces an "AI Usage Index" that compares a location's share of Claude usage with its share of working-age population.
Geography and tasks. The United States accounts for the largest share of overall use. Still, on a per-capita basis, smaller, wealthier countries lead: Israel, Singapore, Australia, New Zealand, and South Korea rank highest. Emerging economies show lower relative usage; India, Indonesia, and Nigeria under-index on the measure. Within the U.S., Washington, D.C., and Utah have the highest usage per capita, ahead of California and New York.
As adoption deepens, use cases diversify. Lower-adoption countries skew toward coding tasks — over half of usage in India — while higher-adoption regions show more activity in education, science, and business functions. The study also finds that high-adoption areas tend to use AI in more collaborative, "augmentation" modes, while lower-adoption areas delegate complete tasks more often.
Shifts in consumer use. Over the past eight months, educational and scientific requests grew as a share of Claude.ai usage, while users increasingly asked the system to complete tasks end-to-end. "Directive" conversations rose to 39% from 27%, alongside a shift from debugging toward new code creation.
Enterprise deployment. Business use via Anthropic's API is more specialized and automation-heavy than consumer use. The study says 77% of API transcripts reflect automation patterns, compared with roughly half on the consumer site. Coding and office/administrative work dominate API use, while education and creative tasks are less common.
Cost plays a limited role in enterprise adoption patterns, the authors say. Higher-cost tasks (measured by token usage) are often used more, suggesting firms prioritize capability and economic value over price. However, the successful deployment of complex tasks appears constrained by access to relevant context and data, pointing to potential bottlenecks in organizations lacking modernized information systems.
Implications. The findings mirror earlier technology rollouts: rapid uptake among early adopters, strong ties to income, and concentrated use in tasks with the best model-task fit.
Anthropic warns that if productivity gains accrue mainly to high-adoption regions and automation-ready sectors, global and regional inequalities could widen.
About the Author
John K. Waters is the editor in chief of a number of Converge360.com sites, with a focus on high-end development, AI and future tech. He's been writing about cutting-edge technologies and culture of Silicon Valley for more than two decades, and he's written more than a dozen books. He also co-scripted the documentary film Silicon Valley: A 100 Year Renaissance, which aired on PBS. He can be reached at [email protected].