Why Won’t Robots Replace Human Workers?
Author: Mark Hillary
The business journals used to say that outsourcing and offshoring was replacing the need for companies to hire people. Now it’s the robots that are supposed to be taking our jobs. Business journals and commentators across the world are suggesting that a wave of automation driven by smart Artificial Intelligence (AI) systems will largely replace the need for workers.
But what is the reality? Forrester Research does believe that a big change is coming, especially to highly developed markets like the US. Their latest study on the global workforce suggested that in 2018, 9% of US jobs will be lost to automation, partly offset by a 2% growth in jobs supporting this automation – the systems need to be managed. The most impacted areas will be back-office and administrative, sales, and call centre employees. A wide range of technologies, from Robotic Process Automation (RPA) and AI to customer self-service and physical robots will change how people are hired and will create a need for different skills.
Analysis by McKinsey is cautious. They warn that just because it is technically feasible for tasks to be automated does not mean that every company will do so. This is similar to the earlier concerns about outsourcing. It is technically feasible to outsource almost every function of a company, yet few companies outsource everything.
McKinsey says that when planning what is possible to automate, first you need to rate jobs and tasks by their technical feasibility – is this a repetitive process that a robot or software bot could perform? For example, when thinking about manual work tasks such as a factory assembly line or food packaging are predictable and it is possible to consider how these tasks could be automated. Construction, forestry, or working with animals is highly unstructured and not predictable and therefore almost impossible to automate.
Let’s consider a simple example. IBA Group created an email bot that sorts incoming emails at the customer support centres. The EmailBot processes typical customer requests, grouping these by content, sending automatic responses, creating tickets, and gathering statistics. However, even a constantly-learning robot cannot process all customer requests. Roughly 50-70 percent of incoming emails are processed automatically and the rest are forwarded to the appropriate employees. Nevertheless, in this case the employees are relieved from repetitive operations and are able to focus on more complicated tasks so value is created.
But factories and manufacturing are just one part of the economy. In most developed economies services are a greater part of the economy. Here there are clear examples of how some automation can be introduced. Look at how customers in McDonald’s are now comfortable using a screen to order their own meal. Amazon has proven that an entire supermarket can be automated, so not even checkouts are required.
Computer Weekly recently published an interesting study of automation that draws the conclusion that the real story is not that robots will cause jobs to vanish. Some jobs will go, but millions will also be created because of the automation. The real story is that many of the jobs we are familiar with today will be transformed.
This rings true. Think about the skills needed by a finance assistant or Human Resources professional in an environment where many business processes will be automated. These office professionals need to be able to control the automation systems and improve them – the back office professionals you are now hiring probably need to be able to code software. That never used to be a requirement in HR, but it will be soon.
The McKinsey research analyses over 800 different types of job and explores the possibility of robots replacing these tasks. You can read the research here, but to my mind it is the transformation of skills that is the real story – not robots replacing workers. Workers need to understand how to work with the robots and control them so they can be more productive.