(Reuters Breakingviews) - Trying to predict how a nascent and promising technology will affect society is hubris, but history suggests people are going to have some serious leisure time if the development of artificial intelligence continues apace. Whether that makes them happy, and how the spoils will be divided, are harder to predict.
Over the past 50 years, technology has tended to grow faster than the wider economy. From 2006 to 2016, the digital economy grew at an average annual rate of 5.6% according to the U.S. Bureau of Economic Analysis, or almost four times faster than the overall output. That sort of expansion appears to be oddly consistent. Revenue earned by technology companies in Fortune’s list of the 100 biggest U.S. firms has, adjusted for inflation, increased at a similar rate for five decades.
American employee productivity has increased by about 2% annually for seven decades. While the higher capital intensity and more skilled labor steadily contribute, what varies more is the ability to deploy technology successfully. Sectors able to automate tasks and reduce workers, such as manufacturing, will generally see higher productivity, while others, such as education, may have a harder time. This process also takes time. In 1987, the economist Robert Solow famously said computers were visible everywhere expect in the productivity statistics. A decade later, productivity shot up.
One outlet for this improved productivity has been largely consistent. Workers tend to trade at least some automation-fueled gains for more leisure time. The number of hours U.S. employees work may not have fallen as fast as in, say, France, but both have declined over the past four decades. Americans have reduced their hours worked annually by 47 hours since 1979, while the French reduced theirs by 274, according to data from the Organisation for Economic Co-operation and Development.
Look back further and the trend is the same. Nobel Prize-winning historian Robert Fogel showed the amount of time people spent on sleep, commuting and meals didn’t change from 1880 to 2000. The hours consumed by housework and employed jobs, however, fell by 3.8 hours per day on average, leaving more time for leisure.
Creating machines that can mimic thought, or think, could unleash economic dislocation on a massive scale. Developed economies have been through several big shifts, though. Around 1800, three-quarters of U.S. workers were employed in agriculture. Farms now employ about 1%, according to U.S. government data. A similar shift happened in manufacturing, where the percentage of total non-farm employment has fallen by more than half since 1979. The adjustment can be wrenching. The loss of manufacturing jobs induced decades of pain on the U.S. Midwest, as it took workers time to find, or retrain for, service jobs. Many older workers never did.
One threat is that the disruption from AI is bigger and faster. For example, about 70% of Americans are employed in the services sector. If technology rapidly makes cashiers, teachers, waiters, and doctors redundant, they may struggle to find other jobs.
A second threat is what the redundant will do. The unemployed found work in the past because there were always things people could do better, or cheaper, than machines. Sure, new technology will create unimaginable new jobs in influencing and caring for people. But if AI systems become far more intelligent than humans, there may be fewer niches for humans to fill.
On the positive side, these trends will probably free up lots of labor time. That doesn’t necessarily mean people will be happy. AI might erode the accomplishment people feel from work, or devalue leisure time because people will have too much of it. Involuntary leisure, through unemployment or underemployment, will be particularly pernicious.
Yet society as a whole should be wealthier. Effective AI means production will be cheaper, goods and services will probably be personalized, new medicines will be developed, and so on. Distributing these gains won’t be easy, as winners will want to keep their earned spoils. Past economic shifts have led to increased economic inequality.
Yet another long trend probably will come into play here. Governments have become steadily larger, and social safety nets more padded, even in the more libertarian United States, as the economy switched from farming to manufacturing, to services.
A richer society will hopefully want to distribute more both for reasons of fairness and social stability. But all this leisure will leave people with lots of time to argue about what level is optimal.