Why heavy Claude users earn neither high nor low wages

An analysis of millions of conversations with Claude.ai reveals who uses AI most and how it correlates with one’s salary.

Unsurprisingly, it’s professionals working within IT and math who use AI the most, and farmers who do not.

Anthropic has just launched the Anthropic Economic Index, which provides a glimpse into how AI is changing the labor market. The initial report is based on millions of anonymized conversations on Claude.ai, an AI assistant built by Anthropic.

According to the report, Claude is mostly tasked with modifying software, debugging code, and troubleshooting networks.

“My game keeps crashing as I only have 8GB of RAM,” is a typical query to Claude. The AI assistant then has the task of modifying the software to improve performance and adapt to new hardware.

Thirty-two percent of Claude queries fall within the “computer and mathematical” category.

But it’s not only developers who rely heavily on AI. The second-largest category of occupations using AI assistants falls under “arts, design, sports, entertainment, and media,” which means they are querying Claude for writing and editing.

“Can you make sure this blog post follows Chicago style,” a typical entry reads. Claude then has to review and rewrite content suitable for that certain publication.

However, it seems that in many cases, AI is not replacing people doing tasks but rather augmenting human work, such as helping users brainstorm, acquire new skills, and double-check their work.

The report also looked into the correlation between AI usage and salary. It turns out that the heaviest users of AI earn neither low salaries nor high wages. Computer programmers and copywriters are in the mid-to-high median salary ranges. Both low-paying jobs (mostly manual labor) and very-high-paying jobs (like obstetricians and gynecologists) have very low rates of AI use.

Anthropic, which will continue to release such insights in the future, said these patterns capture just the beginning of AI’s integration into work.

“As AI systems expand beyond text to handle video, speech, and physical actions through robotics, and as AI agents become more capable of carrying out extended tasks autonomously, the nature of human-AI collaboration is poised to transform dramatically,” the company said.