Economic growth is needed to provide income for workers and to provide tax revenues that support our defense and safety net government programs.  Without Gross Domestic Product (GDP) growth, it is difficult to justify higher wages and it becomes near impossible to fund government at the desired levels.  The advent of robotics and artificial intelligence (AI) led us to hope for their help boosting GDP and wages. So far, we have been dealt disappointment.

GDP growth has been anemic.  In the 70 years between 1947 and 2017, the average GDP growth was 3.22%.  More recently, in the eight years from 2007 through 2014, GDP growth averaged a meagre 1.21%.  In the fourth quarter of 2017, the rate was higher, but just 2.6%.  The chronically poor rate of growth seems closely tied to a miserable rate of labor productivity, which sank to -0.1% in the 4th quarter of 2017.  There does not seem to be a shortage of workers, even with the 4.1% unemployment rate, although small businesses voice the usual complaints of a lack of specialized skills.  There is no surge of people dropping out of the labor force.

Workers leveraged computer and internet technology to boost growth in the 1990 to 2010 period.  We likewise have great hopes that robotics and artificial intelligence will be the golden strain of capital investments that boost labor productivity.  But so far, that new wave of technologies has made little improvement in productivity.

For a couple of decades, industrial robots have been welding cars parts together to produce a chassis, inserting front and rear windows, and rapidly painting the right color and finish on cars.  Such task needing skills and brute force are done without help from the human assembly line workers.  In many respects, the auto-plant workers have been promoted to supervisors, troubleshooters, parts expediters and maintenance support for the robots.  In medicine, some surgeries are better done by robots that are not subject to fatigue or slight tremors in their “hand.”

Artificial intelligence systems have shown high-powered intelligence through success in chess, poker, and the game of Go.  AI systems have helped physicians diagnose disorders, are mastering self-driving for automobiles, and are the human speech interpreters behind the Cortana, Siri and Alexa personal assistants.

Custom made AI is unavailable to help most US workers, but that omission may be cured soon. Developers at Google and Microsoft are working on development of customized AI systems.  That development should sharply reduce the cost of AI systems.  A custom-purposed AI at a low price point is likely to be the winning formula for equipping many more US workers with the productivity-boost of an “AI companion.”

The lack of detectable productivity hikes from AI is surely a consequence of insufficient investment and time for the technology to be integrated into business workflows.  Recent federal tax reform may well provide the economic incentives for businesses install custom AI systems as job companions for selected workers.  The productivity spike they bring may be counted as productivity from capital and labor, but we will be happy to take the win, regardless of how its named.