Why AI Has Not Made Companies More Productive

Summary

The article examines why rapid corporate AI investment has not yet produced comparable gains in economy-wide or company productivity. Gartner projects worldwide AI spending of $2.59 trillion in 2026, while US utilization-adjusted total factor productivity grew only 0.07% over the four quarters ending in the first quarter of 2026; the article also cites a finding that 95% of enterprise generative-AI pilots produced no measurable bottom-line effect. It reviews a University of Pittsburgh working paper that attributes the shortfall to employees resisting AI because they fear job losses, but argues that the study establishes correlation rather than causation and does not show that enthusiastic adoption would have produced larger gains. The article notes that 31% of individual contributors in a Columbia Business School survey expressed enthusiasm about adopting AI, more than half of US workers report using AI at work, and some workers may use AI visibly to demonstrate alignment with company policy. Its alternative explanation is AI overload: one employee can quickly generate more reports, proposals, emails, applications, legal filings, or essays, but other people must spend time checking hallucinations, irrelevant material, and excessive complexity. Examples include job applications per US posting roughly doubling since spring 2022 and AI-generated text appearing in 18% of a sampled set of federal complaints in early 2026, up from 1% in 2023. A thought experiment shows that tripling email output can triple colleagues’ reading burden. The author concludes that workplace AI should be designed around organization-wide productivity and the protection of shared human attention, rather than narrowly maximizing each user’s output.