Houston-Pasadena-The Woodlands, TX · POP 7,904,627 · EMP 3,289,720 · DATA GRADE A
Houston, TX
consensus exposure · range [48.9–49.6] across 5 methodologies
more exposed than 69% of US metros
By methodology
| Index | Exposure |
|---|---|
| OpenAI task exposure (Eloundou et al. 2024) | 49.3 |
| Felten language-modeling AIOE (2023) | 49.1 |
| Microsoft AI applicability (2025) | 49.2 |
| Anthropic Economic Index observed usage (2026) | 48.9 |
| Eisfeldt et al. generative-AI exposure (2024) | 49.6 |
| Consensus (mean) | 49.2 |
Table 1. Where indices disagree, that disagreement is information: prediction-style indices (task ratings) and usage-based indices (observed AI conversations) measure different things. Methods §2.
Secondary measures
payroll-weighted exposure — higher than the headcount number: the exposed jobs are the better-paid ones
of workers are in occupations where ≥50% of tasks are LLM-exposed (Eloundou β; threshold-sensitive — note)
of area employment matched to scored occupations (grade A)
the other side of the ledger: BLS-projected 10-year employment growth (2024–34) for this metro's job mix. National rates × local shares — a mix outlook, not a local forecast. Exposure and growth coexist: the most exposed metros are often also the fastest-growing.
Scenario: if replacement-level AI arrives in 2030
Figure 1. Modeled displacement under the median preset (diffusion k=0.8, ceiling 0.75, automation share 0.45, friction lag 1.5y, attrition 3%/y). Solid: positions eliminated. The gap between gross and layoffs is natural attrition — speed of diffusion, not depth of exposure, determines layoffs. This is a scenario, not a forecast: adjust every assumption.
Where the losses land — and your assumptions
Positions eliminated by 2035 per occupation group, under the arrival year selected above. Drag any multiplier if you think we're wrong about a group — your model, your numbers. Multipliers scale that group's task exposure (×0 = immune, ×2 = double).
Table 3. Group exposure = employment-weighted mean task exposure (Eloundou β over the group's local occupations). Bars use the same scenario engine as Figure 1 (median preset).
Most exposed local occupations
| Occupation | Jobs | Median wage | Exposure [range] | |
|---|---|---|---|---|
| Customer Service Representatives | 65,510 | $40,380 | 91.1 | |
| General and Operations Managers | 97,320 | $119,600 | 56.5 | |
| Retail Salespersons | 78,960 | $31,340 | 67 | |
| Office Clerks, General | 50,040 | $40,370 | 82.2 | |
| Registered Nurses | 65,910 | $99,830 | 48.9 | |
| Secretaries and Administrative Assistants, Except Legal, Medical, and Executive | 33,710 | $46,750 | 89.8 | |
| Cashiers | 54,950 | $29,350 | 54.8 | |
| First-Line Supervisors of Office and Administrative Support Workers | 36,940 | $65,990 | 79.3 | |
| Sales Representatives of Services, Except Advertising, Insurance, Financial Services, and Travel | 28,720 | $63,090 | 94.1 | |
| Project Management Specialists | 35,550 | $99,510 | 71.6 |
Table 2. Ranked by exposure × local employment. Bands on the 0–100 occupation scale.
Compare Houston against any other metro — side by side.