AI Labor Market Tracker: August 2026
What we're reading this month
A brief look at other research on AI and the labor market that informed how we read this month's results.
Iscenko, Z., Strand, S., Chen, Y., et al. (2026). Google's AI & Economy ATLAS v1.0: Mapping Gemini Usage in the Economy. Google & Google DeepMind.
Using 15 million de-identified Gemini interactions, Google finds that AI use is already remarkably broad—touching occupations covering more than 88% of U.S. employment—but still relatively shallow within jobs, covering only about 21% of tasks in the median occupation with observed use; most usage is collaborative rather than end-to-end automation.
Chatterji, A., Holtz, D., Rakholia, N., Tambe, P., & Weeratunga, G. (2026). How Organizations Use AI: Evidence from ChatGPT. Working paper, August 2026.
Linking ChatGPT Enterprise activity to workers and firms, the authors show that adoption varies enormously across companies, but within adopting firms AI use spreads broadly across functions and seniority levels—with particularly intensive use among early-career workers.
Massenkoff, M., Lyubich, E., Sacher, S., Hitzig, Z., Zhang, S., Heller, R., & McCrory, P. (2026). Anthropic Economic Index Report: Cadences. Anthropic, June 26, 2026.
Combining Claude usage data with a survey of users, Anthropic finds that workers believe AI can already perform more of their work than occupation-level exposure measures suggest; more than a third expect AI to be capable of doing most of their work within a year, highlighting how quickly perceived capabilities are changing.
Brynjolfsson, E., Chandar, B., & Chen, R. (2025, revised 2026). Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence. Stanford Digital Economy Lab.
Updated ADP payroll data show no broad AI-driven employment decline, but employment among 22–25-year-olds in highly AI-exposed occupations is now about 19% below that of comparable workers in less-exposed occupations, with the gap driven primarily by reduced hiring and concentrated in occupations where AI appears more substitutive.
Moder, I., & Pommer, T. (2026). AI and the US Labour Market: Effects on Employment Growth. ECB Economic Bulletin, Issue 4/2026.
The ECB finds a growing employment-growth gap between occupations with high and low AI substitution risk—around 15 percentage points between 2019 and 2025—but no corresponding significant difference in wage growth, suggesting that adjustment is showing up first in employment rather than pay.
Babina, T., He, A. X., & Jiang, R. (2026). Canaries in the Gold Mine: Early Productivity Gains from Artificial Intelligence Creating Organization Capital. University of Maryland & NBER.
AI investment was unrelated to productivity growth from 2010–2018, but AI-investing firms have since seen significantly faster growth in sales per worker and TFP over 2018–2024; the authors trace this to firms building organization capital around AI, with gains concentrated among firms that started with the least of it.