About six years ago a study from Oxford University made the startling prediction of how in the next fifteen to twenty years 47% of US jobs are at risk of automation. With only nine to fourteen years left, will half of society’s jobs be stolen by AI? Or rather, is this just a product of our larger fear that “the robots are coming for us and there’s nothing we can do about it”?
They’re not the only ones. Although the figure is surprisingly high, the concern of a shrinking job market has been widely shared. Other reports have estimated that 1.5 million workers in Britain are at a high risk of losing their jobs, while a report from January this year found that a quarter of American jobs are at a high risk of automation by 2050.
To understand why so many are so worried by the prospect of future automation, it is worth looking at technology replacing humans in the past.
The societal displacement caused by the mechanization of agricultural workers in the 19th century is an early example of widespread automation that tends to crop up (pun intended). Despite industrial automation in this field being a process that began in the 18th century, important agricultural advancements such as the steam-powered tractor were made at a critical time.
As a result of the increased productivity and higher yields brought by the mechanical replacements, agricultural workers as a percentage of the labour force continued to shrink throughout the 19th century and into the early 20th.
While many would expect this rise in unemployment to bring forth a plethora of social issues, namely social displacement, structural and long-term unemployment, and a massive reduction in both material and immaterial living standards, mass urbanization occurred instead. These workers then had ample time to transition into the manufacturing sector where they retrained and earned higher real wages.
Throughout the 20th century (and still today), similar fears and warnings have permeated about the deadly potential costs of the loss of manufacturing jobs. Despite this, since 2000 manufacturing employment has fallen by 28%, and there are 22% fewer factories, while during 2018, service sector employment growth accounted for 21% of all employment growth.
There have certainly been transition costs for workers and an in some cases an inability to integrate into a new sector throughout history, but this trend has undeniably shown that the economy can adjust by using the productivity gained by robotic replacements to create new jobs for displaced workers with higher wages and better working conditions.
Yet, a problem remains. After the majority of manufacturing jobs will have given way to automation, computerisation will naturally start (and has already started) to gravitate towards the service industry. This will bring forth a host of concerns, one namely being that company employees will become distanced from their everyday business operations. They won’t have to think about the processes they’re responsible for and will stop thinking independently on how the business can benefit from their labour; thinking is done for them.
As US presidential candidate Andrew Yang put it: “Automation is no longer just a problem for those working in manufacturing. Physical labour was replaced by robots; mental labour is going to be replaced by AI and software.” And it rings true. Employees will become more passive as AI slowly makes more decisions for them. Once this happens, the question must be asked if there is an intrinsic value in a service worker at all.
While this frightening prospect may be assuaged by the history of long periods between each paradigm of automation, the full integration into services could happen sooner than you think. For the technology itself advancing into this field is not concerning, but rather the rate of pace of change of this technology.
Just as the often-cited Moore’s law predicts the exponential growth of processing power, many modern economists are predicting the acceleration of automation to follow a similar path. This process will happen at such a pace that displaced workers won’t have the time to adjust into a new employment sector and retrain their skills, and that’s assuming that there will still be jobs that robots can’t do.
But what about the creative industries that thrive on ingenuity and innovation, I hear you ask? Surely a robot could never replace humans in these fields? Creativity is an attribute that we have almost exclusively reserved for humans throughout history, but examples have continued to materialise of AIs starting to at least imitate it.
IBM’s Watson computer system is an example of one of these AIs. It has already achieved several creative feats such as developing new recipes and food combinations, generating music lyrics, and it even created a movie trailer. Show it to your friends. There won’t be many who can tell the difference between it and your standard preview.
As it turns out, the potential job crisis sentiment is not shared by all. According to a recent study by the OECD puts the figure at a smaller (but still significant) 14%, while researchers at the University of Mannheim have predicted a mere 9% of all jobs are exposed to automation. Whatever the true number turns out to be, and however long it takes, one thing is for certain: don’t be surprised if a robot is sitting at your desk in the morning.