Despite the rapid evolution of artificial intelligence (AI), recent data suggests a troubling parallel to the “productivity paradox” noted by economist and Nobel laureate Robert Solow in 1987. Following the introduction of groundbreaking technologies in the 1960s, productivity growth unexpectedly stalled, leading Solow to remark, “You can see the computer age everywhere but in the productivity statistics.” Today, an analysis of AI’s impact on the workplace echoes this sentiment, as executives express optimism about the technology while broader productivity gains remain elusive.
According to a Financial Times analysis, 374 companies within the S&P 500 referenced AI in their earnings calls, with most reporting favorable outcomes from technology implementation. Yet a study by the National Bureau of Economic Research revealed that among 6,000 executives surveyed across the U.S., U.K., Germany, and Australia, a significant majority perceived little to no impact from AI on their operational efficiencies. Approximately two-thirds of respondents acknowledged using AI, but this amounted to only 1.5 hours each week, and 25% reported no usage at all. Notably, nearly 90% of the firms indicated that AI had not influenced employment or productivity in the past three years.
Executives remain hopeful about AI’s potential, forecasting a 1.4% increase in productivity and a 0.8% rise in output over the next three years, even as expectations of a 0.7% reduction in employment stand in contrast to a projected 0.5% increase in individual employee roles. This mixed outlook raises questions about the technology’s actual benefits compared to its high expectations.
Solow’s Paradox Revisited
As of 2023, research from MIT suggested AI could boost worker performance by nearly 40% compared to those without access to the technology. However, the lack of observed productivity gains has economists questioning when AI will fulfill its promise, especially considering that corporate investments in the technology exceeded $250 billion in 2024. Apollo’s chief economist, Torsten Slok, remarked, “AI is everywhere except in the incoming macroeconomic data,” citing the absence of AI’s influence in employment, productivity, or inflation statistics.
Slok further pointed out that outside the “Magnificent Seven” tech firms, there are scant indicators of AI’s impact on profit margins or earnings expectations. Academic studies provide conflicting insights; for instance, a report from the Federal Reserve Bank of St. Louis noted a 1.9% increase in productivity since the launch of ChatGPT in late 2022, while a 2024 MIT study predicted only a 0.5% increase over the next decade. Daron Acemoglu, a study author and Nobel laureate, acknowledged that while 0.5% growth is preferable to stagnation, it falls short of the industry’s lofty promises.
A notable survey by ManpowerGroup found that among nearly 14,000 workers globally, regular AI usage rose by 13% in 2025, yet confidence in the technology’s utility dropped by 18%, indicating a persistent skepticism about AI’s effectiveness. IBM’s chief human resources officer, Nickle LaMoreaux, highlighted the company’s plans to triple its young hires, suggesting that although AI may automate certain entry-level tasks, this displacement could jeopardize the future leadership pipeline.
Historically, technology adoption often follows a delayed pattern, as seen during the IT boom of the 1970s and ’80s, which ultimately culminated in a significant productivity surge in the 1990s and early 2000s. Erik Brynjolfsson, director of Stanford’s Digital Economy Lab, recently noted that current signs may suggest a similar reversal, as fourth-quarter GDP tracked up 3.7%, despite a downward revision in job gains to 181,000. His own analysis indicated a 2.7% productivity increase in the U.S. last year, attributing it to the transition from AI investment to tangible benefits.
Slok views AI’s trajectory as potentially akin to a “J-curve,” where initial slowdowns in productivity precede exponential growth, contingent on how effectively companies leverage AI technology. Unlike the monopoly pricing seen in the IT sector during the 1980s, the competitive landscape surrounding AI tools has made them more accessible, which could shape the future of productivity gains.
Ultimately, the realization of AI’s potential hinges on enterprises’ willingness to adopt and integrate this technology within their operations. “In other words,” Slok concluded, “the value creation is not the product, but how generative AI is used and implemented in different sectors in the economy.” As stakeholders navigate this uncertain landscape, the lessons of the past may serve as both cautionary tales and guides for the future.
