JOBNATOR

Welcome to the job select screen.

Each character represents a real U.S. occupation. The gauge shows observed AI exposure to its tasks, based on Anthropic’s March 2026 study. It does not count jobs lost or workers replaced.

How to read the gauges

Anthropic combines observed Claude use with estimates of which tasks language models can do, giving more weight to work-related usage and automation. A score of 0% means this method recorded zero weighted exposure for that occupation in this snapshot. The results are based on U.S. occupation definitions and cover the Claude usage studied, so other tools and markets may look different.

Anthropic study · Published dataset