The Anthropic Economic Index (www.anthropic.com)

Anthropic published an Economic Index that maps how people use its Claude model across every US state and hundreds of occupations, letting readers track trending topics and whether interactions are collaborative (augmentative) or fully delegated to the model (directive automation). Key findings include geographically concentrated adoption—AI use is strongest in pockets of knowledge work rather than evenly distributed—and distinctive local usage patterns (e.g., Colorado skews toward travel and event planning, D.C. toward document editing and career advice). Over a nine‑month window, the share of conversations that are directive automation rose from 27% to 39% overall, and reaches 77% among enterprise customers. Anthropic also publishes fuller global data in a companion report. For the AI/ML community this index is significant because it provides empirical, usage‑level signals about where models are deployed, which tasks are being automated, and how real users balance collaboration vs. handoff. Those signals can guide model fine‑tuning, safety evaluation, and UI design for human‑in‑the‑loop workflows, and they highlight socioeconomic implications—regional inequality in access, occupational disruption risk, and enterprise readiness for automation. The rise in directive automation, especially in enterprise settings, underlines the urgency of robust reliability, interpretability and governance work to ensure models perform safely when entrusted with end‑to‑end tasks.