For many of us, the first connotation with automation is an assembly line where a set of robots puts together a car.
In reality, the technical possibilities are far greater. They are most evident in a new trend called Business Process Automation (BPA) that is rapidly gaining traction.
The scope and abstraction of the business processes being automated are steadily increasing to include, for example, sending emails triggered by certain events and entering data into Excel. Sometimes, BPA tools can entirely replace a person using a computer to perform a job involving a repetitive set of well-defined tasks.
BPA is therefore already having a huge impact on the economy. And to best understand this impact you will need think like an economist!
According to a research paper by Oxford University economists Carl Frey and Michael Osborn, it will be possible to automate half of 700 analyzed professions in the USA in the next two decades.
The jobs that may soon disappear all have common characteristics and mainly involve handling uniform objects.
The researchers also observed that computers are very good – oftentimes better than humans – at solving explicitly defined problems and answering specific questions.
Job market automation may therefore sound terrifying, but the truth is more subtle.
According to Frey and Osborn, the American jobs that will survive the BPA boom rely on creativity, social intelligence, or human interaction. And even though computers are superior at answering defined questions, it is still up to humans to handle the definitions and ask the right questions!
Humans are still far superior at those sorts of tasks than even the most advanced machines today. Maximizing efficiency therefore means allocating tasks between humans and robots according to relative strengths, weaknesses, and market conditions.
In other words, just because an activity can be automated doesn’t mean it actually will be.
There are various objective factors to be taken into account like cost and availability of suitably skilled workers.
And many more subjective factors like delicate legal issues, social acceptance, and the actual benefits of increased speed or quality. Most people still prefer a nurse that is human, for example.
Or consider that in highly advanced Huawei factories a fair percentage of assembly lines still rely on manual human labor. Manual labor isn’t cheaper than robotic labor, but humans have superior ability to adapt to changing conditions. Building robots that can assemble a new phone or router model every few months is too hard, if not downright impossible.
A lot depends on how you define your job, of course.
Take another research study, this time from consultancy McKinsey & Company, which used slightly different methodology to analyze automation in the workplace. They defined the specific sub-activities that make up a profession and concluded that about 45% of all compensated activities could be automated using existing technology. But Only 5% of the professions, taken as complete packages of sub-activities, could be entirely automated.
It is therefore safe to say that automation will influence a huge number of professions, but not destroy them altogether. Humans will likely have to cooperate with robots and computers or learn how to operate them.