TL;DR
- Anthropic has shipped a connector that opens the Anthropic Economic Index, its measure of how AI is being used across the economy, directly inside Claude.
- There is nothing to install: open the connectors menu in claude.ai, enable it, and ask questions in plain language about which occupations use AI most or what teachers use it for.
- The useful habit is interrogating a source yourself and asking to see the numbers behind any claim, rather than absorbing a headline or a stranger’s conclusion.
- Anthropic states the Index reflects patterns in Claude usage rather than the whole labour market, and Claude points users back to that caveat.
In plain English
A connector is a switch inside Claude that gives it access to a specific outside data source. Once enabled, questions you type get answered from that dataset instead of from the model’s general training, and you can ask it to show the underlying figures. The Anthropic Economic Index is the dataset in question: Anthropic’s own measurement of how AI is being used across occupations and tasks, including how that has shifted over the past year.
The limit is built into where the data comes from. The Index is drawn from Claude usage, so it describes what Claude users do, and the whole labour market includes work that never touches Claude at all. Reading that boundary is part of using the tool properly.
“Is AI coming for my job?” is usually answered with vibes — a scary headline, a confident stranger on LinkedIn, your own worst 2am guess. This week, Anthropic made it answerable with data instead.
They’ve shipped a connector that opens the Anthropic Economic Index — their measure of how AI is actually being used across the economy — directly inside Claude. There’s nothing to install: open the connectors menu in claude.ai, enable it, and ask questions the way you’d ask a colleague. Which occupations use AI the most? What do teachers actually use it for? What kinds of tasks are people automating, and how has that changed over the past year? You get answers grounded in the data, and you can ask it to show you the numbers behind any claim.
The habit it rewards outlasts the connector. Most people meet a question like this by absorbing someone else’s conclusion. The better move — the one that survives contact with reality — is to go and interrogate the source yourself, then read what it doesn’t cover. Anthropic is upfront that the Index reflects patterns in Claude usage rather than the whole labour market, and Claude will point you back to that caveat as you dig.
That’s the actual skill worth building: not memorising which jobs are “safe,” but knowing how to point an AI at a real dataset and pull a grounded answer out of it.
Replace the vibe with a query. Then check what the query can’t see.