With Amazon Go set to revolutionize the shopping experience in early 2017, many people are worried about its impact on employment. Technological unemployment from automation is something that is of high concern to people in the 21st century, yet it has existed for a long time.
On the one hand, the Luddite fallacy would suggest that despite these changes in technology, people would still be employable as they’ll find new work to do. In it’s simplest form, the Luddite fallacy suggests that jobs aren’t destroyed, it’s the composition of jobs that changes.
On the other hand, however, current trends would suggest that it may become more difficult for people to remain irreplaceable or malleable in their work, both in cognitive and laborious fields.
Automation improves the economy and quality of life
The main argument supporting mass automation is its positive outcomes, with clear evidence suggesting that automation has greatly improved our economy, access to goods and technology, and quality of life.
Findings from a CEP report, titled Robots at Work, showed that in the 17 countries examined, the use of robots led to an increase in both overall factory productivity and wages. Moreover, the increased use of robots raised the annual GDP and labor productivity in each country on average by 0.37% and 0.36% respectively between 1993 and 2007.
Although these numbers may seem quite small, this actually represents a 10% growth in total GDP and a 16% growth in labor productivity over the 14-year time period.
To put these changes in perspective, this had the same economic impact as the introduction of steam technology in Britain from 1850-1910.
The US economy in specific has greatly benefitted from automation. Automation allowed the US economy to remain competitive in the global market as it could compete with nations with lower labor cost like China and India. This moved some manufacturing jobs back to the West. This also allows for a higher task volume for US production compared to nations relying on labor since robots do not incur overtime costs, are not prone to human error, and have no drop-offs in productivity as the day goes on.
Furthermore, automation allows our society to continually innovate upstream.
By continually automating processes and services, businesses are forced to move to the next layer of innovative technologies or risk becoming obsolete. This causes a perceptive shift in how we complete tasks or seek value and is a main driver for advancing societies and technologies.
Automation is holistic, and not only for labour work
Initially, humans developed machines to help automate work done by our own physical labor. Jobs in factories and assembly lines were soon replaced by more accurate and efficient robotic counterparts.
This caused higher unemployment for factory workers, but improved the overall economy and quality of life of society as mentioned above. The newly unemployed had to seek out other opportunities that usually required higher education or training, leading to a more educated and skilled workforce.
Although this was the pattern in the 1940s and 1950s, automation has crossed into the cognitive arena in recent times.
In fact it was estimated that 47% of the employed US population is at risk of losing their jobs to some form of automation or computerization.
Automation does not preclude creative work from its effects either. Neural networks and deep learning are being used to automate even the most creative jobs.
There are different AI applications already analyzing databases of song melodies and lyrics to understand different musical styles to synthesize original music. Two startups of note in this space are Jukedeck and Flow Machines.
nucl.ai is another team working on artificial intelligence specifically for creative applications. They’ve explored ways deep learning can influence the creation of art. So far, their neural net has been able to take simple doodles and transform them into works of art similar to that of famed French impressionist painter Claude Monet.
Although these applications are slow and rife with early stage bugs, this the first step towards making computers think like the human brain.
Automation is collar agnostic and will replace even the most highly skilled jobs, if there is a use case.
Current fields avoiding automation
Although jobs in both the cognitive and physical labour space are being automated, it does not necessarily mean all jobs can be taken over. Based on a US employment survey by the Federal Reserve Bank of St. Louis, it’s not so much the level of cognition that determines if a job will be automated, but the amount of routine involved. Employment for non-routine types of work in both the cognitive and physical labour space have increased significantly since 1983.
Besides non-routine work, jobs requiring empathy or social interaction are the least likely to be replaced. According to a UK Technology Report conducted by Deloitte in 2014, the impact of technological improvement in the UK has caused the fastest growing occupational sectors since 1992 to include: nursing caregivers and assistants (+909%), teaching assistants (+580%), welfare and community workers (+183%), and home caretakers (+168%).
A final thought
The reality of automation likely falls somewhere between the two viewpoints above. The idea that automation will continue to help grow our economies and improve our quality of life is unquestionable, yet we will need to find new and innovative ways to keep people employed and maintain their discretionary income.
All jobs won’t simultaneously be automated, it will be a gradual process that economies and workforces can prepare for. However, we must be cognizant that the rate of automation may soon outpace the rate at which we can move people across jobs.