Anthropic Launches Index to Track AI's Economic Impact

Anthropic has introduced the Anthropic Economic Index, an initiative designed to analyze how artificial intelligence affects labor markets and employment. To build the index's first report, the company used an automated, privacy-preserving analysis tool called Clio to evaluate approximately one million conversations from free and pro users on Claude.ai. Anthropic mapped these interactions to the U.S. Department of Labor's O*NET database, which contains around 20,000 work-related tasks, to determine how the chatbot is being integrated into daily business operations.

The initial findings reveal that AI is primarily used to assist workers rather than replace them, with 57 percent of analyzed tasks categorized as augmentation and 43 percent as automation. The data shows that AI adoption is highly concentrated, with 37.2 percent of Claude queries falling into the computer and mathematical category for tasks like debugging and software modification. The second-largest category, arts, design, sports, entertainment, and media, accounted for 10.3 percent of queries, while physical labor roles like farming, fishing, and forestry represented just 0.1 percent. Overall, about 36 percent of occupations utilize AI for at least a quarter of their tasks, but only 4 percent use it for 75 percent or more of their workload.

The index also highlights a correlation between AI usage and salary. Adoption is highest among mid-to-high wage occupations, such as copywriters and computer programmers, but drops significantly for both the lowest- and highest-paid professions, such as shampooers and obstetricians. Anthropic noted several limitations to the data, including the exclusion of API, Team, and Enterprise tier users, and the fact that Claude's reputation as a coding model might overrepresent technical tasks.

For industry practitioners and researchers, this open-sourced dataset provides a rare, empirical look at actual user behavior instead of relying on subjective surveys. By understanding where Claude is actively driving productivity, developers and enterprise leaders can better identify which workflows are ripe for AI integration and where human-AI collaboration yields the highest economic value.