As someone who has spent over 20 years helping corporations transform their business functions through the implementation of new systems, improved processes and refined roles for their human resources, I find the resurgence of media’s attention to artificial intelligence’s impact on mankind of great interest. Whether you read about drones, driverless cars, or robots being developed for business purposes, it does seem that there are robots everywhere. While industrial and other electromechanical robots are the visual manifestation of our fear of robotics, the developments in artificial intelligence (AI) are not
just limited to robotics. It is this evolution that has driven many to write their perspectives on the effect of robots on our future. As I read these varying viewpoints, I found myself keeping track of who was on who’s side and how fast will it really change the workplace. This article captures my tally of these varying ideas and the
predictions I have for the opportunities and threats of this advancement in technology.
For the purpose of this article when referring to robotics, I am referring to electromechanical robots as well as Robotic Process Automation (RPA). Both of which when enabled by AI may make machines “More Human than Human” which is making people ponder what the future may hold. As these topics are somewhat confusing to most, I will use the following definitions to ground the conversation.
- Robotics is the interdisciplinary branch of engineering and science that includes mechanical engineering, electrical engineering, computer science, and others. Robotics is commonly thought of as the use of computer controlled electromechanical robots that perform manual tasks, especially as it relates to a manufacturing assembly line.
- Robotic process automation (RPA) is the use of a software “robot” to mimic the keystrokes humans make in completing interactions and usage of applications for the purpose of transaction processing, data manipulation and communication across a process. The software robot is just software programming that can replicate either simple activities like data entry or
it can complete more complex cognitive activities when overlaid with artificial intelligence or machine learning.
- Artificial Intelligence (AI) is the use of computer technology to imitate intelligent human behavior, specifically the ability to mimic human decision making processes. A branch of AI study is around machine learning which is allowing machines to have access
to data and to “learn” or adapt to the outcomes of prior activities.
The view of robots and related artificial intelligence technologies ranges from that they are going to either be the downfall of human civilization to that they will cause a renaissance never before seen and cause positive impacts to our economy and lifestyles. In the first camp some notable names including Stephen Hawking and Elon Musk have sounded the alarm that if we do not regulate these new technologies they will surely be the downfall of the human race.
Whether or not you subscribe to either end of the continuum above, most people agree that robots are dramatically changing the way we live and work today and the pace of that change is only anticipated to increase. The Pew Research Center has predicted that robotics and AI will permeate wide segments of daily life by 2025. The items that follow highlight a few of the areas that I believe will be significantly impacted by the adoption of these technologies.
Pew Research shows that 48% of nearly 2,000 experts say that robotics and AI will cause significant impacts in unemployment. 52% of these experts feel that the economy will create as many new jobs as robotics and AI displace. Given that human society suffers from optimism bias, one could view this 50/50 basis to be overly optimistic.
Unemployment rates will rise dramatically
As the cost for technology continues to drop and the abilities of robots and robot software continue to increase, a large swath of blue and white collar jobs will be eliminated. As mentioned in the article by Andrew Yang, Silicon Valley is right—our jobs are already disappearing,
these jobs span a huge range of activities, truck drivers, taxi drivers, manufacturing plant workers, warehouse workers, office clerical staff, auditors, corporate compliance staff, data entry, customer service, telecommunication provisioning staff, banking compliance and risk executives, etc.
While many believe that our economy will be able
to create new markets to absorb the displacement of workers, the convergence of these technologies across a number of industries will at minimum drive a dramatic spike in unemployment and/or underemployment for an extended period of time. Based on the May 2015 job statistics published by the Bureau of Labor Statistics (BLS), an automation rate of less than 15% across the top 40 jobs would displace nearly 42 million jobs. Based upon most experts, 15% is a fairly low estimate so the impact could be much higher. Some see this transition taking place in no less than 10 years but some see it as far out as 50 years from now.
I recognize that these jobs would not be replaced overnight, but I do feel that in the relatively near future these jobs will be affected and likely be displaced over a relatively short period of time. If that period of time was a five-year term, that would make the estimated impact of 8 million jobs a year. We would have to assume some new job creation, but the average across the last 5 presidents is the creation of approximately 1.5M per year. As a result, the impact of this job displacement tidal wave may come quickly and then remain for quite a long time, more than 20 years if the average job creation holds.
Robotics will dramatically change the face of the professional services industry
As mentioned above, the impact of this technological wave will be much broader based than many are predicting. It will affect many white collar jobs as well
as blue collar jobs. In particular the professional services industry will change dramatically. Areas that require large numbers of staff to perform tasks, many of which are routine, such as, audit and certain types of consulting in particular will see a big decline in resource requirements. While this may not be the death toll of the core services for the Big 4 Audit firms, it will result in new models for charging for services that will not be based upon having thousands of staff on the payroll.
Global shipping and logistics will be negatively impacted
As robotics become more intertwined into all manufacturing activities, manufacturing sites will become location independent as labor costs will not drive movement of manufacturing on a global scale. This will result in most products being manufactured closer to home. As a result, the industry of global logistics will be negatively impacted. Global shipping will be scaled down to focus on the shipment of raw materials and inputs into manufacturing eliminating most finished products. The logistics arena will be further impacted by robotic delivery methods from driverless cargo trucks to drone delivery services.
Unfortunately, the local manufacturing will not create enough jobs to offset the impact on global logistics because these local manufacturing “mini-mills” will be fully automated. A current example in the news of the effect on employment is the fact that in early January 2017 Ford Motor Co. decided to scrap their plan for a new plant in Mexico. Instead the company plans to invest
$700M in their operations in Michigan. However, this investment is only anticipated to create 700 jobs,
or one job per $1M invested.
Robotic Process Automation will become commoditized
As RPA technology prices continue to fall and solutions are built for larger companies, those solutions will become pre-packaged and adopted across a broader group of companies in both size and industry focus.
Also the drop in price will allow for a positive ROI on automating functions that use far fewer resources than those that are currently focused on for RPA deployments. This will then truly give rise to what ZDNet has termed
“robot-as-a-service or (RaaS)”. As this happens, cloud based RPA solutions will become prevalent, drive down costs and as a result penetrate the middle market further being used for functions that span smaller and smaller numbers of people.
Social unrest and innovation in social programs will increase
Due to the increasing unemployment rates coupled with the increased momentum around income inequality, there will be an increase in social unrest. This social unrest will drive changes to political thought and social programs that are likely to be permanent. Forward looking countries and organizations are already starting to experiment with “basic income” experiments. The concept is simple, give people a certain monthly stipend with no strings attached and because those people feel that they have
a safety net they will become more innovative and drive the development of new business ideas. Essentially the bet is that people are by their nature more curious/innovative than lazy. This idea was originally brought forward by a small town in Canada in the 1970s and since then has been put into practice in studies in Finland and the Netherlands in 2015 and is was being picked up by Silicon Valley incubator, Y Combinator, in 2016.
While some believe that the adoption of these technologies may not directly result in higher unemployment, at minimum it will likely result in higher unemployment which will put additional strain
on society and drive further changes to the political environment.
RPA will have much larger impacts on the US economy than electromechanical robotics
Most people think of robots as the manufacturing industrial robots seen in car commercials. In fact, the auto industry accounts for nearly 40% of the consumption of industrial or electromechanical robots according to the International Federation of Robotics. This consumption
is only rising as well as the volume of robotics sales to the auto industry has been increasing around 20 percent per year since 2010.
Industrial robotics are dramatically increasing the productivity of a variety of industries beyond the auto industry but the data from the auto industry can be a proxy for other labor intensive industries (e.g. construction, agriculture, food service, mining, fulfillment, etc). Employing a human welder in a factory in the U.S. costs about $25 per hour including benefits, according to a 2015 study by the Boston Consulting Group; that drops to just $8 per hour for a robot, including installation, operating costs and maintenance. By 2030, “the operating cost per hour for a robot doing similar welding tasks could plunge to as little as $2 when improvements
in performance are factored in,” BCG said.
While these electromechanical robots will continue to grow in numbers and uses, the relatively low cost and broad applicability of RPA will quickly outstrip the impact on the US economy. The US economy has been largely services based for a number of decades. The
Department of Commerce published data in 2010 stating that service jobs accounted for more than 80 percent of private-sector employment. The service sector consists of jobs that are well aligned to the capabilities of RPA technologies. Per BLS statistics published in 2015, the six largest components of the service sector State and Local Government, Professional and Business Services, Health Care and Social Assistance, Retail Trade, Leisure and Hospitality, and Financial Activities make up 78%
of the services sector. High proportions of the jobs within these industry sectors will be automated via RPA technologies. Because they drive a bulk of our economy, the impact will be massive even with modest adoption of RPA technologies.
An example of how white-collar jobs will be affected is the use of RPA technology around compliance efforts for banks. Recently, a major US bank was able to implement RPA technology to automate a portion of their compliance efforts. This effort had previously taken 100,000 man-hours per year. With the successful use of RPA, the bank reduced compliance hours to less than 5,000, greater than 95% reduction. In another example an government agency has shown in early proof of concepts that the use of RPA has reduced the man-hours for claims processing by nearly 90%. These type of results are truly transformational and will enable companies to redeploy their efforts towards improved growth and innovation.
While some robotics plays are efficiency based, others are based upon achieving tighter control and eliminating errors. A company that provides pharmacy fulfillment services for long-term care facilities uses electromechanical robotic technology to efficiently pre-package medicines. The robot individually packages the proper dosage for each patient. The technology further aggregates weekly patient specific packages that can be delivered to each floor or wing of long-term care facilities. Without the robotics the cost of pre-packaging would be too high, but the pre-packaging is driven by both control an efficiency. Medicine for each patient is aggregated and then packaged separately by day and time (AM, Noon, PM). These clearly identifiable packages minimize the risk of the provider giving the wrong drugs, providing medicines that may have negative interactions, or giving too much or too little of any one medicine.
Obviously, there is a lot of potential good that these technologies can produce for companies from both efficiency and control but the degree and duration of the impact on the economy are still to be determined. The pace of adoption is the key factor to consider. If the pace of adoption comes relatively soon and over a short period, our current economic and social models will not be able to digest the pace of change as we saw with previous revolutions. If it happens soon we will not be able to create new markets fast enough to absorb the job losses. If it happens over a short duration, the impact will be magnified. The displaced workers from the industrial revolution were farmers who moved to the cities for available jobs. The displaced workers from most other technological advances have been either adopted over a longer period of time or were more heavily focused on one or a few industry segments. As shown above a small level of automation is likely to happen over a range of industries simultaneously.
Some of the robotic technologies, e.g. driverless cars, may be slowed somewhat due to a number of factors including government regulations, ability to drive industry standards and interoperability, etc. We feel that RPA will be much more quickly and widely adopted as both cost and complexity of deploying these solutions is quickly dropping. Once RPA technology is in place we have found that companies will find new ways to extend its use as the deployment of process models takes only a few weeks.My view is that the disruptive power of software automation will drive changes to resource and staffing models beyond manufacturing and into the core white-collar services industry that is such a key driver for the US economy.
About The Author
Patrick Leonard, Managing Director and Founder of the management consulting firm Clarendon Partners, will be publishing a series of robotics white papers that address how robotics can be applied at the enterprise level or within specific functions. His next white paper will focus on the use of robotics in the office of the CFO.
Share this page: