Generative AI: analyzing adoption rates, usage patterns, and productivity effects
Bick et al. examine the adoption of generative artificial intelligence (genAI) in the United States using data from nationally representative surveys conducted in 2024. In particular, they use data from the Real-Time Population Survey (RPS), an online labor market survey of working-age adults (18-64) designed to mirror the Current Population Survey (CPS). The survey asks about genAI use at work and home, frequency and intensity of use, specific products used, and time savings. The main analysis uses combined results from August and November 2024 RPS waves, with over 10,000 respondents. Finally, results are validated using a separate survey, namely the Survey of Working Arrangements and Attitudes (SWAA) and compared to historical data on computer and internet adoption.
- Adoption rates
– 39.6% of respondents reported using genAI either at work or at home.
– 26.5% of workers used genAI at work, with 9.0% using it every workday and 13.9% on some workdays.
– Adoption of genAI has been faster than both personal computers and the internet, relative to each technology’s first mass-market product launch. - Demographic patterns
– GenAI adoption is higher among younger, more educated, and higher-wage workers, similar to early computer adoption patterns.
– Adoption rates vary substantially across occupations and industries, with highest rates in computer/mathematical, management, and business/finance jobs. - Usage and impact
– Between 1-5% of all U.S. work hours involve direct use of genAI.
– GenAI users report an average time savings of 5.4% of work hours, implying a 1.4% time savings across all workers.
– Using a standard model of aggregate production, the authors estimate a potential productivity gain from genAI adoption of around 1.1%. - Task-based analysis
– Actual genAI adoption rates correlate strongly with expert predictions of task-based AI exposure across occupations.
– The most common tasks for which genAI is used include writing (39.5%), administrative tasks (25.6%), and interpreting/translating/summarizing (22.7%).
The study concludes that genAI is being adopted rapidly and broadly across the workforce, with potential for significant productivity gains. However, the authors caution that the long-term impacts remain uncertain and may not be fully reflected in current productivity statistics.
The Rapid Adoption of Generative AI
Authors: Alexander Bick, Adam Blandin, David J. Deming
From: Federal Reserve Bank of St. Louis, Vanderbilt University, Harvard University
