- New data from the US Census Bureau shows that the rate of workers using AI on the job is sharply declining, following three years of booming investment and extensive media coverage.
- According to The Economist, by October 2025, only 11% of workers at large US companies reported using AI for “producing goods and services,” a drop from 12% in the survey conducted two weeks prior.
- In medium-sized businesses (100–249 employees), the percentage of those “not using AI” increased from 74.1% in March to 81.4% – the highest level since the survey began. For large companies (over 250 employees), this figure also rose from 62.4% to 68.6%.
- This is a negative signal for an industry expected to spend $5 trillion on AI infrastructure between now and 2030, as actual revenue lags far behind expenditure.
- Several independent surveys confirm a similar trend: 46% of workers used AI in June 2025, but only 37% by September, according to research by Stanford University. The fintech company Ramp noted that AI usage in US enterprises “peaked” near 40% at the beginning of 2025, then stagnated.
- The reason is attributed to “AI fatigue”: employees feel overwhelmed, while the resulting productivity gains are not commensurate. An EY survey at the end of 2024 showed that over 50% of senior executives admitted to “failing to effectively implement AI.”
- Furthermore, the summer of 2025 saw significant disappointment when models like GPT-5 did not meet expectations, reducing confidence in the new generation of AI.
📌 Summary: The rate of workers using AI on the job in the US is sharply declining. By October 2025, only 11% of workers at large US companies used AI for “producing goods and services.” In medium-sized businesses, the percentage of those “not using AI” increased from 74.1% in March to 81.4%. The reason is attributed to “AI fatigue”: employees feel overwhelmed, while the resulting productivity gains are not commensurate.
