The Geography of AI: Anthropic's Economic Index Tracks AI's Real-World Impact Across 150 Countries

Published:
  Lastmod: 2025-10-20
Random Notes
by Ernest Chiang

Post Title Image (Photo by The New York Public Library on Unsplash)

✳️ tl;dr

  • On 2025-09-15, Anthropic released its third Economic Index (approaching from different dimensions), tracking Claude usage patterns across 150+ countries and all US states for the first time. 1
  • (Possibly the first comprehensive geographic distribution data of AI adoption in the model industry?!)
  • Enterprise API customers show an automation rate of 77%, significantly higher than consumer users’ 50%, indicating that enterprises are actively shifting AI from collaborative tools to productivity replacement solutions.
  • Directive automation jumped from 27% to 39% within 8 months, marking the first time automation (49.1%) surpassed augmentation (47%), reflecting growing user confidence driven by improved model capabilities.
  • API usage shows only 3% price sensitivity (each 1% increase in cost index reduces usage by only 0.29%), with enterprises prioritizing capability and value over cost,
  • The speculated reason is that hidden infrastructure costs far exceed model fees (every $1 in model fees requires an additional $5-10 to deploy and reach production-ready status).
  • Ernest’s field observations align with this: those who complain about token costs typically lack sound organizational operational systems or workflows. Conversely, those who see the overall value created are bold in adopting AI.
  • Approximately 5% of API traffic is dedicated to developing and evaluating AI systems, forming a recursive improvement loop of “AI developing AI,” which is speculated to accelerate capability advancement but also require stronger safety oversight.
  • US interstate GDP elasticity (1.8) is significantly higher than cross-country (0.7), yet income has lower explanatory power, indicating that industry composition and economic structure are stronger adoption drivers.
  • AUI = Anthropic AI Usage Index
  • Washington DC has the highest AUI (3.82), primarily for document editing and information search; California (third) focuses on programming; New York (fourth) prefers financial tasks, with local economic structures directly mapping to AI usage patterns.
  • Educational instruction tasks grew 40% (9% → 13%), scientific research grew 33% (6% → 8%), showing rapid adoption in knowledge-intensive fields, suggesting that high-skilled workers are leveraging AI to enhance professional capabilities.
  • Business management tasks declined 40% (5% → 3%), financial operations tasks halved (6% → 3%), suggesting these fields may be undergoing automation or users are shifting to more specialized tools.
  • Wealthy countries tend to use AI for augmentation, while poorer countries prefer automation, with each 1% increase in population-adjusted usage corresponding to approximately 3% reduction in automation after controlling for task mix.
  • The research uses privacy-preserving classification methods combining the O*NET database (19,498 task descriptions) and Claude’s proprietary classification system for dual verification, ensuring data anonymization.
  • However, its static nature and coarse-grained classification may fail to capture emerging tasks created by AI and programming work of varying complexity. 23
  • The true cost of enterprise AI includes data engineering, security compliance, continuous monitoring, and integration architecture, far exceeding surface-level API fees, which is speculated to explain why enterprises are price-insensitive. 45
  • Model capability improvements (Sonnet 3.6 → 4.x series) directly drive behavioral changes, with better output quality reducing iteration needs, suggesting that future more powerful models may further increase automation ratios and transform human-AI collaboration patterns.

1 4 5 2 6 7 3 8 9

✳️ Knowledge Graph

(Learn more about Knowledge Graphs…)

%%{init: {'theme':'default'}}%%
graph LR
    %% Concept Classes in Orange
    AUI[Anthropic AI Usage Index]:::concept
    TaskAuto[Task Automation]:::concept
    TaskAug[Task Augmentation]:::concept
    GDP[GDP per capita]:::concept
    KnowledgeWork[Knowledge Work]:::concept
    EconDiv[Economic Divergence]:::concept
    DigDiv[Digital Divide]:::concept
    HumanAI[Human-AI Collaboration]:::concept

    %% Instances in Blue
    Claude[Claude LLM]:::instance
    USA[United States]:::instance
    Israel[Israel]:::instance
    Singapore[Singapore]:::instance
    API[API Customers]:::instance
    ConsumerWeb[Claude.ai Users]:::instance
    ONET[O*NET Database]:::instance
    DirectiveAuto[Directive Automation]:::instance
    FeedbackLoop[Feedback Loop]:::instance
    TaskIter[Task Iteration]:::instance
    SoftDev[Software Development]:::instance
    EduInstruct[Educational Instruction]:::instance
    SciResearch[Scientific Research]:::instance

    %% Primary Relationships
    Claude -->|measures usage through| AUI
    AUI -->|correlates positively with| GDP
    GDP -->|explains 70 percent of| AUI
    GDP -->|creates| DigDiv
    DigDiv -->|manifests as| EconDiv

    %% Task Classification
    TaskAuto -->|includes| DirectiveAuto
    TaskAuto -->|includes| FeedbackLoop
    TaskAug -->|includes| TaskIter
    TaskAuto -->|competes with| TaskAug

    %% Geographic Patterns
    USA -->|leads globally in| AUI
    Israel -->|ranks first in| AUI
    Singapore -->|ranks second in| AUI
    USA -->|shows 1.8 elasticity| GDP

    %% User Patterns
    API -->|prefers 77 percent| TaskAuto
    ConsumerWeb -->|balanced 50-50| TaskAuto
    API -->|differs from| ConsumerWeb
    ConsumerWeb -->|increasing use of| DirectiveAuto

    %% Task Categories
    ONET -->|classifies| SoftDev
    ONET -->|classifies| EduInstruct
    ONET -->|classifies| SciResearch
    SoftDev -->|dominates| KnowledgeWork
    EduInstruct -->|grew 40 percent| KnowledgeWork
    SciResearch -->|grew 33 percent| KnowledgeWork

    %% Collaboration Patterns
    HumanAI -->|implements| TaskAug
    HumanAI -->|reduced by| TaskAuto
    DirectiveAuto -->|increased from 27 to 39 percent| TaskAuto

    %% Economic Implications
    KnowledgeWork -->|concentrated in| USA
    TaskAuto -->|may cause| EconDiv
    EconDiv -->|similar to| DigDiv

    %% Feedback Relationships
    Claude -->|enables| HumanAI
    HumanAI -->|measured by| ONET
    TaskAuto -->|reduces demand for| KnowledgeWork
    TaskAug -->|complements| KnowledgeWork

    %% Regional Variations
    USA -->|economic composition drives| KnowledgeWork
    Israel -->|high tech density enables| AUI
    Singapore -->|internet connectivity supports| AUI

    %% Styling
    classDef concept fill:#FF8000,stroke:#CC6600,stroke-width:2px,color:#000
    classDef instance fill:#0080FF,stroke:#0066CC,stroke-width:2px,color:#FFF

✳️ Further Reading

  1. https://www.anthropic.com/research/economic-index-geography ↩︎ ↩︎

  2. https://www.pewresearch.org/social-trends/2023/07/26/2023-ai-and-jobs-methodology-for-onet-analysis/ ↩︎ ↩︎

  3. https://academic.oup.com/pnasnexus/article/3/9/pgae320/7758639 ↩︎ ↩︎

  4. https://www.anthropic.com/research/anthropic-economic-index-september-2025-report ↩︎ ↩︎

  5. https://www.pymnts.com/artificial-intelligence-2/2025/enterprises-confront-the-real-price-tag-of-ai-deployment ↩︎ ↩︎

  6. https://www.aalpha.net/blog/advantages-of-software-development-outsourcing-to-india/ ↩︎

  7. https://www.yourteaminindia.com/blog/outsourcing-in-india ↩︎

  8. https://serjhenrique.com/which-economic-tasks-are-performed-with-ai-evidence-from-millions-of-claude-conversations/ ↩︎

  9. https://www.statista.com/outlook/tmo/it-services/it-outsourcing/india ↩︎