Anthropic wants to know what AI is doing to your job

By the time official government employment reports flag a shift in the labor market, rank-and-file workers have usually spent months quietly handing their day jobs over to AI. Traditional economic indicators show what already happened yesterday, while token counts only tell you how hard servers are sweating. Neither reveals what happens inside actual offices.

Anthropic wants to fix that blind spot. The company's Economic Research team just rolled out the Anthropic Economic Index Survey, an ongoing monthly study designed to track how AI is rewiring real-world work before aggregate labor data catches on.

Interviewing the workforce through Claude

The study runs through Anthropic Interviewer, the company's automated conversational research tool. It builds directly on an earlier pilot from December 2024, when Anthropic gathered 81,000 open-ended responses to analyze worker sentiment.

Instead of checking dry multiple-choice boxes, participants will answer dynamic questions about their current workflows. The study asks which specific tasks people are offloading to Claude, whether they are seeing genuine productivity gains, and if their organizations are already changing hiring plans or restructuring roles. It also probes expectations for the next twelve months and asks what people hope an AI-driven economy will look like in ten years.

To keep data fresh, Anthropic plans to pull a small, randomized sample each month from active users whose personal accounts are at least two weeks old. If selected, you will spot an in-app banner on claude.ai or the Cowork desktop app, or receive an email if you primarily stick to mobile. Anthropic says responses will be combined with privacy-preserved usage data to produce future research briefs and refine its own models.

Why it matters

Anthropic is basically turning its own user base into an early-warning radar for the global labor market. We think this is a clever move because the gap between deploying AI and seeing it register in macroeconomic statistics can span quarters, if not years. By tapping people at the exact moment they use the tool, Anthropic can spot labor shifts as they happen rather than doing an autopsy after mass layoffs occur.

For workers and managers, these upcoming reports could offer a much more honest view of office automation than corporate white papers. If you want to know which tasks are genuinely vanishing versus which ones AI still fails at, first-hand accounts from power users are arguably the best signal available.

That said, there is a clear catch. Claude users are naturally tech-forward early adopters, meaning this survey samples the people most likely to benefit from AI rather than those displaced by it from the outside. If Anthropic treats these findings as a proxy for the entire economy, we think it risks painting a much smoother transition than what mainstream workers are actually experiencing.