If you have been keeping up with the news lately, automation and AI are all the rage. Advancements in AI are happening at a staggering pace and it’s helping small and large companies automate things that were once done by humans. Whether we are talking about automating extremely repeatable tasks like medical claims coding with Robotic Process Automation or physical robots that do order fulfillment in warehouses, there is an abundance of use cases where automation could be deployed.
The question that always motivates me to dig deeper is this; why are companies taking something like AI and automation so seriously now? Back in 2011, I can remember sitting at my desk after I had just automated a process. This process took me a half day to complete without automation, and I went from actively performing the process to simply monitoring the process. I couldn’t fathom why companies weren’t doing this for every process possible because of the sheer impact it had on my own job. So why did it take until now for companies to realize what I did seven years ago? The answer really lies in the inherent nature of capitalism and the fierce competition for growth and resources.
Capitalism = Endless Growth
Capitalism requires endless growth. Public companies, in particular, are required to provide shareholders value. They do this in the form of profits and growth. This ever-expanding growth that powers business needs something important for it to run; resources. Historically, business has been powered by people. Companies hire people to do work and that work results in something tangible from which the company can profit. If you haven’t noticed, there are about one million articles on boosting productivity of your employees. This is because the more productive your employees are, the more profit you can generate (in theory). The problem with any resource is that it is finite. In the business context, people are finite. There are only so many people in the world. There are only so many hours in a day that people can work and be productive. With this finite resource, there is only one option following the traditional business model, and that’s to hire more people. What happens when there aren’t enough people to hire? McKinsey Global Institute and many academic papers suggest that we are approaching a scenario where our growth targets will be too large for the number of people in the global workforce. For the global economy to continue growing, automation and AI are not a choice, they are necessary to secure our collective future.
Hold on — I Thought the Robots Were Coming for Our Jobs?
Let’s take a step back before we start panicking about the robots taking over. The impact that AI will have on our society could be very similar to what occurred in the Industrial Revolution. Prior to the late 1700s, people were manufacturing products by hand in their homes. The Industrial Revolution brought a wave of mass production factories that used special machines to make products. The changes that came from the first wave of the Industrial Revolution (1790–1850) caused a bifurcation in society and social class.
At the start of this revolution, progress was slow and improvements to quality of life were nearly non-existent. There were almost no laws or regulations for this new manufacturing era in which roughly 80 percent of the workforce would be considered working class. This led to terrible things like child labor, poor living conditions, high unemployment (some estimates as high at 8 percent) and decreased quality of life for many. One of the primary reasons for this low quality of life was the lack of laws and regulation and the primary beneficiaries of a lawless society are always people with money and power. Once new laws took hold, workers could really start to see a benefit. Real wages nearly doubled between 1810 and 1850 and the overall quality of life for most people was elevated. Anti-trust laws and minimum wage could also be directly attributed to the Industrial Revolution due to some of the pessimistic views of the impact that capitalism had on society at that time. What exactly does this mean for AI?
Yes, automation and AI technologies will be replacing jobs. Just like the Industrial Revolution, the first ones to go are unskilled labor that are easily automated due to their repetitive nature. This is a chilling prospect for a lot of people; the uncertainty of what your future will hold as it relates to employment. As Svetlana Sicular at Gartner put it “Many significant innovations in the past have been associated with a transition period of temporary job loss, followed by recovery.” Those in unskilled careers that have high automation potential will be the ones hurt the most. The long-term implications to society are where the benefits are.
Even with the temporary short-term job losses, Gartner predicts that AI will create more jobs than they will replace by 2020. Just like the Industrial Revolution, people will have to learn new skills and new laws will need to dictate how society can curb some of the negative side effects as AI becomes more prevalent in everyday life. The real benefit is the impact that AI will have on the quality of life for humans, from digital assistants to advances in medical treatment that can spot disease more accurately than humans can. Looking at the impact to the Chicago job market can shed some light on what the rest of the US might look like.
The Impact AI and Automation Will Have in Chicago
First and foremost, no one knows the true impact automation will have on our workforce. As Donald Rumsfeld best put it, “We don’t know what we don’t know.” Frey & Osborne estimates that 47% of the jobs in the United States are at risk to be automated. That is a high number that has the potential to truly alter the course of what being “employed” in the United States means. One source isn’t good enough when you’re looking at such a complicated problem. Arntz, Gregory, and Zierahn came to a much different conclusion. They estimated that automation will only effect around 9% of jobs in all OECD countries. Now clearly the scope of the second study is much larger and could skew the results for the United States, but also there is some disparity here. So, riding on the heels of a bunch of research, lets dive into what might happen in Chicago.
Kellogg School of Management did some great research to understand the economic impact of automation broken down by city. The researchers knew that each city would be affected differently, and automation could have a profoundly different effect on cities that vary in size. They wanted to develop ‘risk of computerization’ estimates to provide an educated guess on the relative impact automation will have on each city. Based on the makeup of the occupations that reside within the Greater Chicago Area, they estimate nearly 60% of jobs, or 3.8 million, will be impacted by automation. Let’s break this number down so we can truly understand what this means for the job market in Chicago.
There are several things to call out about this estimate of “impact.” First, this impact calculation is trying to understand how easy it is for a computer to replace a “skill” as outlined by the US Bureau of Labor Statistics dataset. Since there are hardly any jobs that consist of a single skill, the only true way of understanding the uncertainty of complex jobs, like technology consultants, is to calculate the entropy. For those unfamiliar with entropy, it plays a key role in Information Theory. It refers to disorder or uncertainty. If a job in this study had a low entropy, there is a good chance that the job didn’t have many skills and it’s reasonable to say that job will be automated. The higher the entropy, the more uncertainty there is therefore the probability of automation goes down. The second thing to understand about this scary number is that neither job creation due to automation or reallocation of skills was factored into this study. As said previously, it is predicted that job creation will outpace job loss in the very near future. Maybe breaking down an occupation might help shed some light on what might really happen in the coming years.
Robotic Process Automation and Bookkeeping, Accounting, and Auditing Clerks
Robotic Process Automation is ideal for the automation of extremely repeatable tasks done on a computer. If there is a team of people auditing insurance claims, often they are approaching audits from a very consistent rules-based mindset. If you can extract that tacit knowledge from an auditor and clearly define what those business rules are, Robotic Process Automation has the potential of automating the entire audit process (or at least part of it). When I say rules based, think of IFTTT and the way you create simple and clearly defined rules for when an applet triggers an action. Robotic Process Automation is nearly the same. Keeping the number of business rules low increases your chances of a more successful automation. Since I have personally been a part of this scenario of applying Robotic Process Automation to audits, I know the impact to the job market will be significant if Robotic Process Automation continues to grow like it has been. Looking at the research around these types of jobs confirms my suspicions.
Using Kellogg’s interactive tool built on the research they performed that was previously cited, we can see the impact to people in professions such auditors, accountants, and clerks is almost 100%. Based on my experience working with companies to deliver automation technologies, I believe this number to be completely correct. But once again we must really focus on the term “impact” and understand that this is an overly simplistic view on how AI technologies will really change society and the workplace.
At this point, it’s understood that various AI technologies will have a significant impact on our society. After reviewing what occurred during the Industrial Revolution, we can be certain that something similar will likely take place. We know from history, and from what AI does, that the first jobs to go are the ones that are simple and repeatable jobs and tasks that are currently performed by people. After truly breaking down that research, we can be confident that the jobs being replaced are only one very small facet of what AI is really going to change for our society. The short-term job losses and lack of laws and regulations should be a big concern to all and be addressed at the same pace as technology is advancing. Between the job creation, reallocation of skills, and the improved quality of life that much of society will experience as a side effect of AI, the long-term benefits to society will bring humanity into a new era that will truly lift up all of society across the globe. We’re still at the very beginning of this new era.