Recent speeches by Federal Reserve Bank officials and members of the Board of Governors suggest that productivity is on the minds of policymakers. One common theme is that artificial intelligence could increase productivity.

First, let’s define productivity.

Economists use two alternative definitions of productivity—labor productivity and total factor productivity (TFP)—that are somewhat correlated but not necessarily equivalent.

Our FRED graph above shows both of these measures, annually, for the period 1988-2024. Labor productivity is calculated as real GDP divided by hours worked. This can vary with changes in technology and in the amount of capital per worker. TFP, on the other hand, is computed by effectively holding capital per worker fixed and more directly represents the effect of technology on productivity. One thing to note, though, is that changes in labor productivity encompass changes in TFP.

How could AI impact productivity?

Artificial intelligence is a technology that can improve efficiency and increase productivity capacity. So, it has the potential to increase both labor productivity and TFP.

Some economists believe that a portion of this productivity boost might have already been realized but that most of the gains are likely yet to come, as AI adoption widens. AI’s effect on these two measures of productivity could also vary. For example, construction of data centers—now and in the future—increase capital. This could lead to an increase in labor productivity, but it has an uncertain effect on TFP, depending on whether output rises more than the capital stock. Read more on this topic in the St. Louis Fed On the Economy Blog.

How this graph was created: Search FRED for and select “Private Nonfarm Business Sector: Total Factor Productivity.” Click on the “Edit Graph” button and select the “Add Line” tab to search for and add “Private Nonfarm Business Sector: Labor Productivity.”

Suggested by Brooke Hathhorn and Michael T. Owyang.