Individual workers using AI are producing more output per unit of labor, but total economy-wide productivity growth has stalled, mirroring the early Internet era's 'productivity paradox'

According to a London School of Economics study, employees using AI are more likely to produce the same amount of work in less time, potentially saving an entire workday a week. However, a Harvard Business Review study of 200 employees at a U.S. technology company found that time saved on tasks was often redirected into other work, resulting in more time on the job and higher risk of burnout.

Labor productivity (output per unit of labor) has posted solid gains in recent years, but total factor productivity (TFP)—a broader metric measuring how efficiently the entire economy converts inputs into output—has struggled to post significant growth since a post-pandemic surge. This divergence suggests individual workers are working faster while the workforce as a whole has not necessarily become more efficient.

A March study by the Atlanta Fed surveying around 750 corporate executives found productivity is improving thanks to AI, but perceived productivity gains reported by executives were larger than what researchers could actually measure from indicators such as company revenue. The Fed attributed this gap to 'delayed output realizations.'

The San Francisco Federal Reserve Bank noted that the current AI-economy dynamic mirrors the early and mid-1990s, when massive IT investments initially failed to translate to measurable efficiency gains—a period economists called the 'productivity paradox.' The researchers wrote: 'If today mirrors what we experienced in the mid-1990s, we may be in the early stages of a productivity boom driven by AI that will only become clear in retrospect.'