Not so long ago the idea of robots patrolling neighborhoods or caring for children was the domain of science fiction. While robots have yet to replace police and daycare workers, technology has become so advanced that automated systems are taking on greater roles in society and the workplace.

Academic research has explored the diverse impacts of technology on employment — what happens when jobs shift elsewhere or when they’re atomized through Internet-enabled technologies. A 2015 study from Uppsala University and the London School of Economics looked the economic impact of industrial robot use in 17 countries from 1993 to 2007 and found that robots contributed to the economy, partly by helping humans do their work better.

A 2015 study for the National Bureau of Economic Research, “Robots Are Us: Some Economics of Human Replacement,” uses an economic model to explore the potential impact of increased workplace automation. The authors — Seth Benzell, Laurence Kotlikoff and Guillermo LaGarda of Boston University and Jeffrey Sachs of Columbia University — develop a model that calculated the initial and final states of an economy with two inputs to production (capital and code) and two types of workers (low-tech and high-tech). They then tested the impact that variations on workplace conditions and industrial policies will have on the economy.

Key findings of the study include:

• Increased workplace automation could produce both “economic misery” and prosperity. Specifically, three consequences were found to be highly probable: “A long-run decline in labor share of income (which appears underway in OECD members), tech-booms followed by tech-busts, and a growing dependency of current output on past software investment.”
• As technology improves and its use in the workplace expands, the demand for high-tech workers falls. At the end of the simulation, nearly 68% of high-tech workers end up in the service sector, earning approximately 14% less than they did previously.
• As high-tech workers return to the service sector, the wages of low-tech workers rise 41%, then fall to 17% above the initial steady state wage — higher than the initial state, but lower than during the “boom.” In effect, the drop in high-tech worker compensation generates a boom-bust in low-tech worker compensation.
• In the long run, national income increases in the short term, but then falls by 17%.
• Adding a “positive tech shock” — a technical innovation that increases reduces costs or increases productivity — to the model causes a 13% short-term increase in national income, but national income then falls again by 28%, ending up lower than in the initial steady state.
• During a positive tech shock, labor’s share of national income also rises in the short-term but then falls, from 75 to 57%.
• The positive tech shock also causes consumption of goods to decrease by 28%, and the price of services to decrease by 43% as compared to before the technological breakthrough.
• Some public policy options were found to reduce the negative long-term impacts on workers and the economy. For example, a high national saving rate mitigates the impacts of a positive tech shock, resulting in workers earning very near their initial steady state wages (rather than far less), but able to consume 20% more with those wages than they were before the shock.
• Some policies to mitigate the negative effects were found to be likely to backfire, including requiring that all code be open source or restricting the labor supply — these solutions were found to further hurt wages, savings and capital stock.

The authors conclude: “Our simple model illustrates the range of things that smart machines can do for us and to us. Its central message is disturbing. Absent appropriate fiscal policy that redistributes from winners to losers, smart machines can mean long-term misery for all.”

 

Keywords: technology, artificial intelligence, AI, robotics