AI's impact on the US labor market Report Interpretation

Overview

This US macro dashboard examines whether AI adoption is already damaging employment. BofA concludes that current data do not show broad AI-driven job destruction, although recent college graduates and selected white-collar sectors warrant attention.

Core views

BofA’s central premise is that technological shocks primarily replace tasks rather than entire occupations or aggregate labor demand. The report expects an initial displacement effect to be followed by a reinstatement effect as new tasks and roles emerge. This matters because AI targets the white-collar, service-providing economy, which accounts for nearly 84% of US private-sector jobs; nevertheless, the institution remains optimistic that AI will not produce the widely feared employment collapse. The report first tests the relationship between AI exposure and employment across 206 industries using the Felten, Raj and Seamans (2021) AI-exposure index. Industries with the highest exposure have seen employment remain broadly unchanged since the November 2022 launch of ChatGPT 3.5, while the least exposed industries have grown by around 2%. But the more granular cross-industry evidence shows little correlation between AI exposure and job growth. BofA therefore argues that factors other than AI—such as excessive hiring in the initial years after 2019—may help explain relatively weak hiring in highly exposed sectors. The same conclusion holds for aggregate hours worked: firms do not appear to be reducing the intensive margin of labor in sectors with greater AI exposure. The institution then examines whether AI is weakening labor demand rather than directly reducing employment. It defines labor demand as employment plus job openings and uses the Census Bureau’s Business Trend and Outlook Survey measure of AI use in the prior two weeks. Since January 2026, there has been little correlation between AI-use levels and changes in labor demand. The exceptions BofA flags are information and finance & insurance: both have high AI use and experienced declines in labor demand over the five months ending in June, potentially indicating efforts to offset labor costs through AI adoption. Youth labor outcomes are the main area of concern. Unemployment rates for people aged 22–27, including college graduates in that age group, have risen from their 2023 lows and sit above 2019 levels. BofA attributes part of last year’s rise to trade-policy uncertainty delaying hiring, but notes that recent college graduates have seen little improvement even as uncertainty faded. In the institution’s view, AI may therefore be contributing to weaker entry-level outcomes as businesses invest in capex in pursuit of lower labor costs over time. BofA also identifies an offset through the AI investment buildout. Data-center-related capex has supported a reacceleration in nonresidential construction employment, while manufacturing industries tied to the buildout have grown faster than those that are not. Nonresidential construction added 95k jobs year to date and AI-related manufacturing added 32k; together, these sectors represent about 25% of new private-sector jobs this year. With capex plans continuing to be revised higher, the report expects these sectors to continue helping offset potential AI-related job displacement in the near term. Overall, BofA finds few signs that AI is currently a major headwind to broad job growth, outside the modest rise in youth unemployment.

Analysis framework

The report compares employment growth and hours worked across 206 industries ranked by an AI-exposure index, then compares AI-use rates with changes in labor demand, defined as employment plus job openings. It separately examines youth unemployment and capex-linked construction and manufacturing hiring to assess offsetting effects.

Methodology notes

• Industry AnalysisVolume-price decomposition

  Employment and hours-worked comparison across industries with different AI exposure

  The report separates employment levels from hours worked to test whether AI exposure is linked to either fewer jobs or lower labor intensity.

• Other

  Cross-sectional comparison using the Felten, Raj and Seamans AI-exposure index and Census Bureau AI-usage data

  BofA compares outcomes across industries with differing measured AI exposure or usage rather than treating a simple timing overlap as proof of causation.

Key data

• Service-sector share of US private-sector jobsnearly 84%The report notes that AI targets the white-collar, service-providing economy.
• Industries analyzed206Cross-industry employment-growth analysis using the AI-exposure index.
• Employment in highest AI-exposure industrieslargely unchanged since November 2022Measured since the launch of ChatGPT 3.5.
• Employment in least AI-exposed industriesaround 2% growthComparison with highly exposed industries.
• Information sector AI usage and labor-demand change42.1%; -1.9%AI usage in June 2026; change in job openings plus employment from January to June 2026.
• Finance and insurance AI usage and labor-demand change34.8%; -1.1%AI usage in June 2026; change in job openings plus employment from January to June 2026.
• Nonresidential construction jobs added95k year to dateAttributed to the AI investment and data-center buildout.
• AI-related manufacturing jobs added32k year to dateCapex-linked job growth.
• Share of new private-sector jobs~25%Combined share from nonresidential construction and AI-related manufacturing this year.

Impact & implications

BofA’s reading of the data supports its view that AI has not yet become a broad job-destruction force. Capex-driven hiring in goods-producing sectors is currently an important offset, but the report highlights entry-level labor outcomes and selected high-AI-use white-collar industries as areas where adverse effects may emerge.

Risks

• AI may be contributing to persistently weaker employment outcomes and higher unemployment for recent college graduates aged 22–27.
• Information and finance & insurance may be using AI to offset labor costs, as both showed declining labor demand over the five months through June.

What to watch

• Labor-demand trends in information and finance & insurance, which BofA identifies as sectors to watch.
• Whether unemployment among recent college graduates aged 22–27 improves as trade-policy uncertainty fades.
• Whether higher AI-related capex plans continue to sustain construction and manufacturing hiring.