A new report on the use of the AI model Claude reveals that its adoption is uneven across geographies and industries, with users increasingly entrusting it with automated tasks. According to the third Anthropic Economic Index, these early patterns are beginning to reshape work and the economy.
The analysis finds that AI use correlates strongly with income. People in higher-income countries and US states use Claude more per capita. However, the composition of local economies also plays a significant role. For example, Claude is used disproportionately for finance-related tasks in New York and for tourism-related requests in Hawaii.
Globally, Israel, Singapore, and Australia lead in per capita adoption. The report also notes a surprising trend: users in countries with lower adoption rates tend to use Claude for automation more frequently, while users in high-adoption countries prefer a more collaborative approach.
Over the past nine months, the share of “directive” conversations, where the AI works with minimal human input, has risen sharply from 27% to 39%. This suggests users’ confidence in AI is growing rapidly. Businesses using Anthropic’s API automate tasks far more often than individual consumers, with 77% of their interactions showing automation patterns. This points toward potentially significant future impacts on the labor market.