Why Even Economists Aren’t Safe From Automation
During President Obama’s farewell address he stated that “the next wave of economic dislocations won’t come from overseas. It will come from the relentless pace of automation that makes a lot of good middle class jobs obsolete.” He then advocated for a new social compact for future generations of Americans that would ensure them the education they would need to be able to participate in the labor market. Just last week, I wrote a post regarding whether a strong or weak dollar was better for the economy, and in particular the industry which President Trump wants to resurrect—American manufacturing. However, regardless of whether or not he can weaken the dollar to make American exports more competitive and in turn increase manufacturing jobs, developed economies are moving towards automation. So what exactly is automation, and how will it reshape our job market?
Automation is the process by which machines, with or without artificial intelligence (AI), perform jobs, tasks, and operations formerly held by humans. Automation without AI could be something simple like an assembly line machine that constructs a car without any human intervention. AI could potentially be like something in HBO’s Westworld with AIs resembling humans. However, current AIs look more like computers, or rather like nothing as they could simply be a software program. Some commercial examples of AIs include: facial recognition software, or voice recognition on your phone. The AI field is constantly evolving and getting more complex, however we’re still far from having synthetic humans take our jobs.
The automation that is more likely (and to some extent is already happening), is the automation of basic manufacturing and service jobs. The most obvious is perhaps in retail and the fast food industry. At most big grocery stores in the US, there is the ubiquitous (and infamously faulty) self-checkout option. This past year McDonald’s implemented its new kiosk system. Essentially, there are a set of touch screen kiosks near the front counter where you place your order. This simple touch screen interface, full of images of what your food (should) look like makes it simple and easy for consumers to place their orders. Furthermore, there can be more kiosks than there could ever be registers, and people to staff those registers. While the role of the cashier has become automated at some McDonald’s, the preparation of the food is still conducted by humans.
In addition, this past December, Amazon revealed its newest project titled “Amazon Go,” which is essentially a store without any floor employees (there will be a manager who oversees the operations). Here, cameras monitor shoppers and record what they take off the shelves and as they leave their Amazon account is charged accordingly. Theoretically, there could be no human worker on the store floor and the store would operate as normal. This further shows the trend of retail towards automation.
Just this past Tuesday, Uber announced that its set of 16 self-driving cars will begin to pick up passengers in Arizona. Uber, since its genesis, has hit the taxi industry hard as prices via Uber tend to be typically lower than that of a traditional taxi or car service. Many Uber drivers have found the job as a means of income support, or as a full time job. However, that may change within the next 10 years. Self-driving cars are only getting better as technology evolves, and while these set of 16 cars will definitely not put Uber drivers out of a job immediately, it will place pressure on them.
So automation appears to only be affecting manufacturing and basic services, right? Not quite. Machines and computers are becoming better at doing more than just manual tasks. Medical technology and software is in some cases better than humans at discerning malignant from benevolent tumors. In finance, high-frequency trading (HTF) has become fairly popular. Essentially, it’s a program that uses advanced trading algorithms and high processing power to track and analyze markets. With this information, HTFs execute trades when a set of criteria is met. While these trades aren’t completely automated, as they do require humans to develop strategies, they nonetheless represent a portion of the industry that has become automated. Furthermore, there is a push within the AI-PhD and hedge fund communities to optimize performance by using “machine learning” to completely automate the fund. While we are by no means here, this may represent the future we are headed to. One where an entire hedge fund is operated by a set of computers that can learn from changes in the market.
Even jobs that required a person to be highly skilled and educated could become computerized. For example they found that the probability was for the job of an economist to become computerized was a 0.43—a position that typically requires both a PhD and expertise in the field.
The two examples above represent a small portion of industries that will be restructured by increased levels of automation. Any industry that has opportunity for data analytics and decision making based on that information is at risk of being automated. Even highly technical fields like finance and medicine will be impacted by automation. In Carl Frey and Michael Osborne’s piece, The Future of Employment: How Susceptible are Jobs to Computerisation?, they created a scale from 0 to 1 with 0 being no chance of computerization and 1 being a certain chance. Even jobs that required a person to be highly skilled and educated could become computerized. For example they found that the probability was for the job of an economist to become computerized was a 0.43—a position that typically requires both a PhD and expertise in the field. The professions that had the lowest probability were typically professions where individuals explicitly worked with people (teachers, therapists, human resource managers), professions that required human creativity (choreographers and pattern makers), and the people who would build and maintain the automated machines.
Our economy is moving away from human labor and into machine labor, whether it be in manufacturing, basic services, or more advanced services. Computers and algorithms may begin to fully replace human beings in the future. At that point what happens to our job market? What happens to the level of unemployment? And how does the economy continue to grow, with the risk of destroying disposable income?