Half of Claude users say AI can already handle half their work according to Anthropic survey

The survey population matters enormously here: nearly 10,000 Claude users are, by definition, people already committed enough to the tool to be surveyed by its maker. Asking active users whether AI handles half their work is roughly equivalent to asking gym members whether exercise is effective, and Anthropic has a direct commercial interest in the answer.

This is largely disconnected from recent activity in our archive, so there is no prior Modelwire thread to pull forward. The story belongs to a broader category of vendor-sponsored workforce research that has proliferated across the industry, where Microsoft, Google, and now Anthropic periodically release survey data that tends to validate their own products' centrality to the future of work. That pattern is worth naming, because it shapes how the underlying anxiety statistics get framed and distributed.

Watch whether an independent labor research group (Pew, NBER, or similar) publishes a comparable methodology on general worker populations within the next six months. If their task-delegation numbers land anywhere near 50 percent, Anthropic's figures gain credibility. If they land well below 20 percent, this survey is better read as a product marketing document than a workforce signal.