AI Has Passed the Bottom of the Productivity J-Curve, Stanford Expert Says

Artificial intelligence has moved past the deepest trough of the "productivity J-curve," according to Erik Brynjolfsson, director of the Center for Digital Economy at Stanford University. Brynjolfsson asserts that the global economy is now entering a steep upward phase of productivity growth.

He explains that the introduction of general-purpose technologies often creates a paradox where productivity appears to stagnate or decline. This occurs because companies must invest heavily in intangible assets—such as restructuring organizational workflows and retraining employees—which masks actual technological progress in national statistics. In the case of AI, these early investment and learning costs created a temporary dip in measured efficiency.

The Signals of Recovery

To determine when a technology has passed this low point, Brynjolfsson identifies four critical signals: productivity growth within AI-intensive industries, confirmed growth in company production and sales, the launch of entirely new products or services, and massive labor relocation.

According to data from the United States, the first and fourth signals are already appearing. Industry-specific growth in AI-heavy sectors is becoming evident, and significant shifts in labor allocation are underway, suggesting the transition toward the upward slope of the curve has begun.

A Faster Transition

One of the most striking aspects of the AI transition is its projected speed. Historically, the adjustment period for general-purpose technologies has been lengthy; for example, Enterprise Resource Planning (ERP) systems took approximately 15 years to fully manifest their productivity gains. Brynjolfsson expects the AI adjustment period to be significantly shorter, potentially taking only three to five years.

This accelerated timeline suggests that the benefits of AI could hit the broader economy much faster than previous digital revolutions. However, these gains are not guaranteed for every firm. Brynjolfsson emphasizes that productivity surges will only be realized by companies that actively redesign their decision-making methods and work responsibilities to fit the AI era, rather than simply layering new tools over obsolete processes.

The Human Element

As the economy shifts toward this high-growth phase, the role of the human worker remains a central point of contention. Brynjolfsson recently led a statement signed by more than 200 economists, including 16 Nobel laureates, urging that AI be used to complement human labor rather than replace it.

While the data suggests an explosive entrance into a new era of productivity, the ultimate impact will depend on how labor is relocated and how the workforce is integrated with these new systems. Observers will now be watching for the remaining two signals—broad sales growth and the emergence of entirely new AI-native product categories—to confirm the full ascent of the J-curve.