Today’s automobile manufacturing plants are manned by highspeed robots that lift steel panels, hold them in place, weld them together and present the assembly to the next specialized robot. Down the line, a robot slowly spins the entire steel body of a car so a “paint” robot can spray it uniformly – flat panels and crevasses. Nimble robots thread wire harnesses into the engine compartment, cabin, and trunk, connecting the instruments, sensors, controls and lights. A handful of human inspectors watch for anything that seems amiss, ready to halt the process, diagnose the problem, and make technical adjustments.
Today’s auto manufacturing is nothing like the thousand-person shifts, conveyor belts and compressed-air hand tools from the 1950s. Today’s productivity is much higher, a consequence of deep investments in automation using hydraulic robots and computers. The relatively few humans on today’s assembly floor have more specialized skills than their predecessors. Other factories went through a similar transformation. Technology has already displaced many of the US manufacturing jobs, and the capability of robots, especially those powered by artificial intelligence, will soon replace many more. Self-driving cars (per Elon Musk) are the obvious example.
Further development of artificial intelligence (AI) and robotics may offer successively higher levels of productivity due to automation in law, medicine, and architecture. These may be attractive investments, but they leave us with difficult questions. Will there be enough new jobs created to absorb the workers displaced by automation? Can displaced workers acquire the right skill mix to fit those new jobs and can those new jobs pay enough keep families afloat?
If US families had deeper insulation against disruption in earnings, the consequences of AI and robotics would be less threatening. As expected, the degree of unpreparedness is felt mostly in lower income families. They have modest bank savings, few expect pensions, and they have low balances in retirement contribution plans such as 401(k)s and IRAs. Offsetting their meager assets are $12.6 trillion in household debts. Debt leaves most of us with a low net worth.
In 2013, the bottom half of the income strata who were a decade away from retirement (age 53 through 62) held a median balance of just $23,000 in their 401(k) or IRA. Vulnerability is not limited to the poor. That same age bracket but in the 50th to 95th income percentile held a median of just $130,000. If workers from this age bracket suffer unemployment due to technological displacement, they probably have insufficient time to be retrained, land a suitable job and bolster their retirement nest egg. The economy is unlikely to begin offering yields that are meaningful from such small IRAs. Sixty-nine percent of Americans have less than $1,000 in their savings accounts, and the average unemployed family would quickly deplete the median $60,000 in home equity. The family of typical, displaced middle-aged workers can look forward to an economically bleak future, with $1,341 per month from social security as their primary income.
The job displacement problem is acknowledged among those who understand AI and robotics. To offset the tax revenue losses from displaced workers, Bill Gates recently suggested that robots be assessed a tax equivalent to what would have been collected from human workers. Others are chipping the tarnish off socialist schemes such as a guaranteed annual income. Regardless of how alarming these proposals might seem, we will need to address any groundswell of unemployment due to AI and robots. We especially need to limit the cost that our compassion would impose on those still employed.
AI and robotics will create both real problems and real opportunities. Now is not too early to consider how we will manage those challenges.