UBS economist Arend Kapteyn contends that rising U.S. productivity isn't driven by widespread AI deployment. He argues current adoption levels are far too low to justify claims AI is the primary driver, stating output growth stems mainly from tech capex and consumption by the top income quintile. The bank's analysis points to Kansas City Fed research suggesting only 2% of productivity gains can be attributed to AI adoption variations across industries.

UBS's own survey reveals just 17% of firms implement AI at scale, up only modestly from 14% last year, with agentic AI lagging at 5%. Respondents consistently overestimate implementation speed. Obstacles include unclear ROI, regulatory concerns, and a lack of expertise.

With over 80% of firms yet to integrate AI into production and minimal capex outlays, UBS deems it implausible to interpret productivity increases as evidence of widespread AI implementation. Meaningful gains remain in the 'figuring it out' phase.