Quick Buzz Feed
Mapping federal data collection through the lens of increased adoption of AI in workplaces across the nation.
Federal data on unemployment, wages, and job availability in the U.S. labor market are fragmented, often not timely, and difficult to link. Measuring artificial intelligence adoption by individual firms is challenging due to inconsistent, binary surveys unconnected to worker outcomes. Existing federal data cannot conclusively demonstrate how AI is changing the labor market. Therefore, improving federal data collection infrastructure is crucial for policymakers to understand AI's impact and design effective, timely responses to ensure equitable economic growth.
It is difficult to predict AI's full impact on the U.S. labor market, including new opportunities, job erosion, and worker displacement. Better data is essential for informed policy, program design, and research. Currently, there's a lack of detailed, real-time, and impartial federal data on how AI affects workers, firms, and the economy. This brief is the first in a series to map the federal data infrastructure regarding AI's impact on unemployment, wages, and job availability, aiming to inform future policy recommendations.
Understanding AI's role in job losses requires precise unemployment data, which is challenging to obtain. Sources like the Current Population Survey (CPS) provide a monthly rate but exclude discouraged workers and offer limited occupational detail. Unemployment Insurance (UI) claims are timely but miss many ineligible or unaware workers. Longitudinal Employer-Household Dynamics (LEHD) data links workers to employers over time but has a lag and lacks occupational details. The WARN Act data is limited in scope. These sources struggle to isolate AI's impact due to various influencing factors and data limitations.
Tracking worker compensation and job quality is vital to understand AI's impact on income inequality. Various federal surveys like the CPS, Current Employment Statistics (CES), National Compensation Survey, and National Longitudinal Surveys offer insights into earnings, but each has limitations such as self-reporting, lack of demographic or occupational detail, exclusion of independent contractors, or data lags. The IRS provides tax filing data with independent contractor income, but also with delays. Connecting these diverse and often inconsistent data sources to paint a comprehensive picture of AI's effect on wages and earnings remains a significant challenge.
Policymakers and workers need timely information on job availability as AI reshapes the nature of work. The National Labor Exchange (NLx) offers real-time online job postings but misses informal networks. The Job Openings and Labor Turnover Survey (JOLTS) provides industry-level data but lacks occupational specifics. Other BLS and Census Bureau surveys detail employment by industry, occupation, and demographics but no single source captures all dimensions in real-time. The O*NET Program describes job tasks, but its infrequent updates limit its use for tracking rapid AI-driven changes. Overall, federal data struggles to track how AI specifically alters job demand in real time.
Measuring AI adoption by firms and its impact on workers is difficult. While NLx job postings offer indirect signals of AI-related skills, federal surveys like the Census Bureau’s Business Trends and Outlook Survey (BTOS) and Annual Business Survey (ABS) provide binary or inconsistent data on AI use. A critical gap is the lack of linkage between firm-level AI adoption data and worker-level outcomes such as employment or wages. International surveys and privately funded initiatives attempt to fill this void, but federal collection is crucial for comprehensive and impartial understanding.
The central question of how AI adoption affects worker employment, compensation, and job opportunities cannot be answered by any single data source, federal or otherwise, due to significant data gaps. This recognized challenge has led to calls for improved federal data infrastructure from advisory councils and plans for initiatives like the AI Workforce Research Hub. Proposed bipartisan legislation also aims to address these data collection needs, setting the stage for future recommendations to enhance federal data on AI and its impact on the workforce.