The adoption of AI is accelerating across the global economy, but its use remains uneven and geographically concentrated. Anthropic’s latest economic index report, New Building Blocks for Understanding AI Use, analyzed the real-world usage of its AI model, Claude.

The report looked at how AI, specifically Claude, is currently being used, the types of tasks where it delivers the greatest productivity gains, and how adoption differs across countries and occupations. Anthropic’s report reinforces that that the economic benefits of AI are contingent upon adoption, which varies widely across industries, skill levels, and economic contexts.

Using a framework of five “economic primitives”, Anthropic looked at the use and impact of AI by assessing task complexity, skill level, purpose, AI autonomy, and success. Using these metrics to assess the economic impacts of AI, the report found that AI delivers the largest productivity gains for tasks that require higher levels of human capital. As a result, it sees adoption rates the highest amongst white-collar professions. The tasks best supported by Claude are those that require more education compared to the overall economy-wide average, suggesting AI is augmenting higher-skilled work rather than replacing it.

In 2025, Anthropic’s adoption data found roughly 36 per cent of jobs used Claude for at least a quarter of their tasks. Compared to one year later, the figure has grown to 49 per cent of jobs. AI tools are increasingly supporting a growing share of everyday work across occupations and enhancing productivity.

However, AI adoption and use vary significantly across countries. Countries with higher GDP per capita are adopting AI more widely and tend to use the technology for work, personal productivity, and leisure. Those leading in Claude usage include the U.S., India, Japan, the U.K., and South Korea. In contrast, lower-income countries are not adopting AI at the same rate and tend to use the technology more for education, coursework, and learning support.

Artificial intelligence has the potential to transform economies, enhance productivity, and become the next industrial revolution. Earlier research from Anthropic estimated that widespread AI adoption could increase U.S. labour productivity growth by 1.8 percentage points annually over the next decade, roughly double the current rate. When the probability that AI will complete a task successfully is considered, the number falls to 1.2 percentage points, and 1.0 percentage points for more challenging tasks. Even on the lower end of these rates, AI still has the potential to see the U.S. return to growth rates not seen since the late 1990s and early 2000s.

Despite its impact on economies and productivity, global AI adoption remains uneven, with higher-income economies adopting the technology at a faster pace. As the world enters a period of rapid technological change, Anthropic’s report provides data and analysis to assess the real-world impacts of AI adoption on workers, industries, and regions at large.

With the launch of The Anthropic Institute, a new initiative focused on researching the economic, legal, and societal impacts of AI, the AI Adoption Initiative looks forward to continued insights from Anthropic. Its research will help inform the technological transition of economies and support the responsible adoption of AI.

Find Anthropic’s full report on the New Building Blocks for Understanding AI Use here.