The AI Waterline: every state filled to the share of its workers in the 26 jobs most exposed to AI. Washington's tide runs highest, nearly three times Mississippi's, while businesses' own AI use keeps rising.
AI exposure by state: share of each state's workers employed in the 26 occupations most exposed to AI. Each square is one state, filled from the bottom (a full square = 6%). The dashed line in every square marks the U.S. rate, 3.23%. States above it are highlighted; click the legend to switch.
Workers in the 26 occupations Virginia Works rates most exposed to AI, divided by each state's total employment, using BLS occupational employment for May 2025 (SmartAsset's calculation). The U.S. rate is derived: the 51 state counts summed (5,024,380) over the implied state totals (each count ÷ its share). Exposure measures contact between AI and job tasks, not forecast job losses.
AI-exposed jobs by state, top 10 by number of workers in the 26 high-exposure occupations, May 2025 employment
Big states lead on headcount even where the share is middling: California is 4th by share (3.977%), New York 18th and Florida 21st. Tooltip shares are of the 5,024,380 summed across all 51 rows (derived). SmartAsset's Texas figure is printed as exactly 500,000.
Jobs most exposed to AI: share of each occupation's tasks that AI could augment or replace, Virginia Works scores for 18 of the 26 occupations
Whole-percent labels as printed on Virginia Works' employment chart (p. 6); names shortened from the BLS titles (e.g. Translators = Interpreters and Translators, Transcriptionists = Medical Transcriptionists, Office support = Office and Administrative Support Workers, All Other). The other 8 of the 26 (proofreaders, correspondence clerks, court reporters, media and communication workers, telephone operators, communications equipment operators, statistical assistants and CNC tool programmers) appear on its p. 5 ranking without printed values. Software developers are highlighted because they are the largest high-exposure occupation in Virginia's employment.
AI adoption by U.S. businesses: share that used AI in any business function in the prior two weeks, by survey wave
Census changed the question in November 2025 from AI use “in producing goods or services” to use “in any of its business functions,” so this series starts with the new wording and is not comparable with earlier BTOS rates. The current questionnaire calls it “Super Intelligence or SI (formerly referred to as Artificial Intelligence or AI).” Biweekly survey estimates; small moves between waves may be noise.
AI exposure vs. AI adoption by state: share of workers in high-exposure jobs (left to right) and share of businesses using AI (up), dot size = exposed workers
Business AI use is the simple average of the six latest BTOS waves (reference periods Jul 27 to Oct 18, 2026), our calculation, because single-wave state estimates swing widely. The correlation across all 51 is about 0.55 (0.49 without DC), our calculation. The two measures differ: one counts workers by occupation, the other counts businesses.
AI job losses so far: how businesses say their AI use changed total headcount over the last six months, Oct 5–18, 2026 wave
First wave with this question, so there is no trend yet. A companion question on new hires: 3.3% said AI decreased hiring and 2.5% said it increased hiring. Census's download does not say whether these questions went to all businesses or only AI users; shares are as published.