Anthropic lets readers explore its Economic Index with Claude
Anthropic shipped a product surface that lets users ask Claude about the Anthropic Economic Index, turning labor research into an interactive culture beat.
Labor research used to live in PDFs. On July 22, Anthropic turned its own dataset into a chat surface. The lab announced “Ask Claude about the Anthropic Economic Index,” a connector that lets anyone query Index data inside Claude. The same day it published a research agenda for a $200 million Economic Futures Research Fund. Culture coverage cares because a frontier lab is packaging work-and-automation findings as something ordinary users can interrogate — on the lab’s own product.
What the Economic Index ask feature does
The Anthropic Economic Index measures how AI is actually used in the economy, grounded in Claude usage patterns. The new connector, available in claude.ai’s connectors directory with no install, lets users ask questions such as which occupations use AI most, how people in a given state use Claude, what tasks teachers automate, and how automation patterns changed over the past year. Answers are supposed to be grounded in Index data, with Claude pointing back to sources and limitations.
That is the product claim. The culture claim underneath is bigger: economic anxiety about AI should be queryable by a teacher or a journalist in the same interface they already use for drafts. Anthropic is not waiting for a Brookings digest to mediate the numbers.
The Economic Futures Research Fund agenda
Same day, Anthropic committed $200 million to external research on interventions that might make economies more flexible as AI scales. The agenda prioritizes five areas: shaping AI’s impact on workers inside firms, equipping people for transitions, modernizing income support for displacement, building worker stakes in AI-driven growth before disruption, and generating evidence on public investments. Grants are aimed large — primarily $5–30 million, nothing under $1 million from this fund — for universities, research institutes, and nonprofits that can run ambitious pilots.
Read together with the Index connector, the day is a two-surface play: interactive data for the public, big external bets for researchers. Both keep Anthropic at the center of how AI’s labor story gets told.
Why interactive labor research is a culture beat
When a lab lets you chat with its employment dataset, it is competing with unions, newspapers, and academic centers to set the metaphors of disruption. “Which jobs use AI most?” is not a neutral UX prompt. It is a framing device that starts from observed Claude usage, not from pink slips or guild headcounts. That can still be useful. It can also quietly redefine “impact” as “what our users already do,” which is a different question from “who lost the contract.”
Mid-2026 culture fights over Hollywood animation, synthetic performers, and civic AI gifts all need numbers. A conversational Index gives Anthropic a friendly path into those fights without publishing another static chart alone.
Limits of chatting with the lab’s own index
Anthropic’s own caveat matters: the Index reflects Claude usage, not the labor market as a whole. Chatting with the lab’s model about the lab’s usage data is not the same as auditing displacement. It will not tell you whether Katzenberg’s 90% warning is right, or whether Netflix’s 300-title AI tally cut specific crafts. It will tell you what Claude users appear to automate.
Treat the connector as a public research surface with a conflict of interest that is disclosed by structure. Use it. Then check the free datasets, independent surveys, and payroll realities the Index does not claim to cover. Watch whether journalists cite Claude’s Index answers as if they were BLS series. That is the failure mode. The success mode is faster public literacy about AI-at-work patterns — labeled as usage data, not destiny.
