Key Findings
The AI Transition in American Work — 4 Findings
What the data says about where the US stands in the AI transition, which workers are most exposed, and where the labor market signal is visible.
Section 1 — GAIA Behavioral Data
USA in the Global AI Adoption Index
How US AI usage patterns compare to Brazil, India, Germany, and the global average. Data from Anthropic's Claude.ai behavioral logs, February 2026.
Use Case Breakdown — US vs. Peers
Collaboration Style — US vs. Peers
GAIA Scores — US vs. Peers
Data access by request. Cite as: Aghabarari (2026) DOI: 10.5281/zenodo.20320112
Section 2 — AI Exposure by Occupation
Most and Least AI-Exposed US Occupations
AI exposure scores from Eloundou et al. (Science 2024) — share of occupation tasks performable by GPT-4 augmented with tools (E1+E2 specification). Click any occupation to see all scores.
Top 20 by AI Exposure (E1+E2)
Click an occupation →
Select an occupation from the chart to see its full GAIA profile.
Bottom 20 by AI Exposure (E1+E2)
Click an occupation →
Select an occupation from the chart to see its full GAIA profile.
Data access by request. Cite as: Aghabarari (2026) DOI: 10.5281/zenodo.20320112
→ Full Occupation ExplorerSection 3 — AI Exposure & Wages
AI Exposure vs. Wages: Who Is Most at Risk?
Across all 923 US occupations, higher AI exposure is concentrated in higher-wage jobs — but a pocket of high-exposure, low-wage roles faces the starkest displacement risk. Wages shown are FRED industry-level averages (Apr 2026) mapped to each occupation's sector.
GAIA-E Exposure Score vs. Industry Wage — all 923 occupations
Data access by request. Cite as: Aghabarari (2026) DOI: 10.5281/zenodo.20320112
Each dot is one occupation. Wages are FRED sector averages (Apr 2026) assigned by occupational group — a proxy for within-occupation wage variation. Hover for details.
Section 5 — FRED · Labor Market by Industry
Wages, Job Openings & Unemployment — Before and After ChatGPT
Three labor market dimensions by industry from the St. Louis Fed FRED database, 2020–2026. The vertical line marks November 2022. High-AI-exposure industries (Information, Prof/Business) show distinct earnings and openings trajectories.
Average Hourly Earnings by Industry ($/hr, monthly 2020–2026)
Job Openings by Industry (thousands, JOLTS monthly 2020–2026)
Unemployment Rate by Industry (%, monthly 2020–2026)
Pre vs. Post ChatGPT — Average Hourly Earnings by Industry
Section 6 — BLS JOLTS · Labor Market Trends
Hiring, Quits & Layoffs — Before and After ChatGPT
Monthly Job Openings and Labor Turnover Survey data from the BLS API, 2020–2026. The vertical line marks November 2022 — the ChatGPT public release. The key question: did AI uncertainty change worker and firm behavior?
Total Nonfarm — Hires, Quits & Layoffs (Monthly Rate, 2020–2026)
Data access by request. Cite as: Aghabarari (2026) DOI: 10.5281/zenodo.20320112
Industry Comparison — Average Quit Rate Pre vs. Post ChatGPT
Section 7 — O*NET Task Content
What Do High-Exposure Workers Actually Do?
O*NET task statements for high and low AI-exposure occupations. Understanding the task content helps explain why some occupations score high. Data: O*NET 29.0, merged with GAIA-E scores.
Select an occupation to see its tasks:
← Select an occupation
Data access by request. Cite as: Aghabarari (2026) DOI: 10.5281/zenodo.20320112
Section 9 — Google Trends · ChatGPT Awareness
ChatGPT Search Interest in the US — 2022 to 2026
Weekly Google Trends index for "ChatGPT" in the United States. Index: 100 = peak interest. The awareness curve shows continuous acceleration since the November 2022 launch — the US has not yet hit a plateau.
Weekly "ChatGPT" Search Interest — United States (Jan 2022 – May 2026)
Data access by request. Cite as: Aghabarari (2026) DOI: 10.5281/zenodo.20320112
ChatGPT Interest by US State (cumulative, Nov 2022 – May 2026)
Data access by request. Cite as: Aghabarari (2026) DOI: 10.5281/zenodo.20320112