Economic Primitives
Economic primitives are simple, foundational measurements of AI use that help estimate AI’s economic impact beyond task labels alone.
Key points
- Anthropic’s January 2026 Economic Index adds five primitive categories: task complexity, human and AI skills, use case, AI autonomy, and task success [src-069, src-070].
- The primitives are generated by asking Claude classifier questions about anonymized Claude.ai and first-party API transcripts [src-069, src-070].
- They extend the prior automation/augmentation measurement by distinguishing dimensions that can otherwise be conflated. A directive translation request can be high automation but low autonomy because it requires little decision-making [src-069].
- Anthropic treats the classifiers as directionally accurate rather than exact. Their value is in combined patterns across tasks, regions, occupations, and platforms [src-069].
- The primitives enable richer questions: where AI is reliable, how long real tasks are, what education level prompts and outputs require, and which occupations have effective task coverage [src-069, src-070].
- Anthropic presents primitives as a leading indicator: they can track whether AI use becomes more reliable, more autonomous, more business-critical, or more concentrated in particular occupational tasks over time [src-070].
- The March 2026 report uses primitives longitudinally: Claude.ai prompts became slightly less complex on average, required less estimated human time, and were given more AI autonomy as the user base broadened [src-071].
- Primitives also become inputs into learning-curve analysis: higher-tenure users show higher human-input education levels, less personal use, more collaboration, and higher success [src-071].
Related entities
Related concepts
- Anthropic AI Usage Index
- Real-World AI Task Horizons
- Effective AI Job Coverage
- Task-Level Deskilling and Upskilling
- Augmentation-Automation Perception Gap
- AI Adoption Learning Curves
- API Workflow Migration
Source references
- [src-069] Anthropic – “Anthropic Economic Index report: Economic primitives” (2026-01-15)
- [src-070] Anthropic – “Anthropic Economic Index: New building blocks for understanding AI use” (2026-01-15)
- [src-071] Anthropic – “Anthropic Economic Index report: Learning curves” (2026-03-24)
Recommended next
Keep reading from this thread
From 491 indexed pages and articles.
- Wiki concept Anthropic Economic Index Anthropic's privacy-preserving research program for measuring how Claude is used across tasks, geographies, platforms, and economic contexts. Related by primitives
- Wiki concept Effective AI Job Coverage Estimates the share of a worker's time-weighted duties that AI can successfully perform, rather than only counting which Related by primitives
- Insight AI Measurement and Experimentation How to measure AI product impact with evals, adoption metrics, online experiments, guardrails, and cost tracking Readers have engaged with this next