4 July 2026

The political geography of AI exposure

The Brookings Institution | Mark Muro, Todd Jones, and Shriya Methkupally

Sixty-two of the 100 most AI-exposed U.S. counties voted Democratic in the 2024 presidential election, indicating that potential political anxiety related to artificial intelligence job disruption is concentrated in blue-leaning areas. This correlation, reflecting occupational sorting, suggests workers in these counties, like New York and San Francisco, could become flashpoints for AI-related economic concern in upcoming midterm elections.

Research leveraging Anthropic's Claude model usage data reveals that high AI exposure aligns with high-level office or information-based work, reinforcing an urban focus. States such as Massachusetts and California exhibit 13-17% AI exposure and voted blue, while states like Nevada and Wyoming show 10-11% exposure and voted red. Swing states like Arizona and Georgia also display very high AI exposure levels. This pattern implies that blue places, concentrating workers in occupations susceptible to AI-driven job dislocation, may foster agitated voters in the AI era.

No comments: