Markets
CROSS-CATEGORY: The AI Productivity Paradox Crystallizes — Three Independent Surveys Converge on Gap Between Executive Claims and Measured Reality
Three data sources published within 72 hours paint a consistent picture: (1) Duke/Fed NBER paper shows CFOs claim 1.8% AI productivity gains but revenue data shows much smaller actual impact; (2) the same survey finds 502,000 AI job cuts planned (9x increase) despite modest real productivity; (3) Fortune's analysis shows Fed data of 2.2 hours/week time savings has not translated to measurable revenue gains. The convergence across different research teams (Duke, Federal Reserve Banks of Richmond/Atlanta/St. Louis) using different methodologies (executive surveys, revenue analysis, time-use studies) suggests the gap is structural, not anecdotal — and mirrors Solow's 1987 computer productivity paradox.
↳ Follow the thread