More than half of U.S. workers now use artificial intelligence on the job, and many say it meaningfully trims their workloads, according to new U.S. Census Bureau survey data. About 55% of respondents reported using AI tools at work; roughly one-third of them said the technology saves one to two hours per task. Smaller shares cited larger gains: 15% saved three hours and another 15% saved four hours. Employees most often deploy AI to search for information or technical help (37%), draft text (32%), generate ideas (32%), summarize or translate information (31%), and handle administrative tasks (27%). Fewer use it for data analysis (21%), training (16%), customer support (12%), coding, medical uses, or logistics. Economists caution that adoption often follows a “J-curve”: productivity can dip as organizations invest in and learn new systems before longer-term gains materialize. MIT research in manufacturing underscores the pattern, with initial slowdowns preceding improvements. The Census findings, drawn from a weekly national survey of households, sketch a snapshot rather than a definitive productivity scorecard, but they highlight how quickly AI is permeating everyday work and where employees see the most immediate returns.
Related articles:
Generative AI at Work
The Productivity J-Curve: How Intangibles Complement General Purpose Technologies
The Future of Jobs Report 2023
