The AI Investment Paradox: Billions Spent, Productivity Gains Elusive

A stark disconnect has emerged in corporate boardrooms. Despite massive investment and relentless discussion, artificial intelligence has so far failed to move the needle on productivity or employment for most companies. A February 2026 study from the National Bureau of Economic Research, surveying nearly 6,000 executives across four major economies, found nearly 90% reported no measurable change in either metric over the past three years. While two-thirds of firms say they use AI, the average weekly use sits at just 1.5 hours; a quarter of leaders don’t use it at all.

The findings echo a famous observation from the 1980s, when economist Robert Solow noted computers were visible everywhere except in productivity statistics. Today’s data paints a similar picture. Stanford economist Nick Bloom, a co-author of the NBER study, notes senior executives report using AI for less than an hour a week, with 28% using it zero hours. The net effect on jobs has been negligible. A separate PwC study in April 2026 suggests a widening gap: 74% of the financial value from AI is captured by just 20% of companies. These leading organizations are distinguished by their focus on overhauling core processes and building strong data foundations, while others remain stuck in a cycle of pilot projects.

Expectations, however, remain high. Executives forecast a 1.4% productivity lift and a slight reduction in jobs over the next three years. Workers are more skeptical, with surveys showing many save little to no time using AI tools their bosses champion. The implementation challenge appears profound. As Bloom summarizes, 'As of yet, there has not been a big effect.'

The parallel to the slow-burn impact of information technology offers a note of historical context. Widespread IT adoption eventually fueled a mid-1990s productivity surge. Whether AI follows a similar 'J-curve' depends on moving beyond experimentation to genuine operational integration. For now, the promise of AI is held in check by the practice of it.

Source: Webpronews