Data shows first recorded dip since 2022
Companies have been rolling out AI at speed across boardrooms and back offices, but fresh survey data hints that the enthusiasm is beginning to taper off.
Figures from the US Census Bureau show large companies are pulling back on their use of artificial intelligence after a rapid build-up in spending.
Among firms with more than 250 staff, adoption rates peaked at just under 14% earlier this summer before slipping back to about 12%. The fall is modest but telling, given that these enterprises have thrown billions into datacentres and infrastructure to push AI projects forward.
Torsten Sløk, chief economist at Apollo Academy, remarked in a blog post that “the bottom line is that the biweekly Census data is starting to show a slowdown in AI adoption for large companies.”
The Census Bureau, which surveys 1.2 million firms biweekly, has been the US government’s main yardstick for tracking AI use since 2022.
When the survey began, fewer than 4% of firms said they used AI in production. That figure had risen above 5% by mid-2024, reflecting a steady climb before this year’s dip among bigger players.
Early overreach and workforce impact
This development suggests many companies may have rushed into AI rollouts in hopes of cutting costs, only to find the payoff slow to come through.
A Stanford study linked AI adoption to a 13% drop in jobs for younger US workers, while in the UK, a good number of young adults are afraid they will lose their jobs to AI.
In many cases, firms cut back on staff but failed to bring down expenses, leaving them frustrated with the technology’s limited impact on profit.
It is not just the biggest corporations that are beginning to backpedal on AI adoption. A survey of 1,500 smaller firms in June also pointed to falling usage. The data highlights the blurred line between leaning on AI to support staff and swapping workers out entirely. Businesses still need people to steer these systems, but as more routine tasks are handed over to algorithms, job losses mount.
For now, the momentum remains uneven, with some sectors pressing ahead on AI while others ease off as early hype gives way to more cautious assessments.