The most common concern about automation is whether or not it will take jobs away from people.
Certainly, automation in some forms has already had an impact on the job market and the economy. In industries such as manufacturing, some jobs formerly done by people are now being done by machines. As McKinsey points out, the potential for automation within any particular job depends on a number of factors including the nature of the work, technical feasibility, and cost.
Despite anxieties about the effects of automation on employment, history and data paint an overall optimistic, but complicated, picture. A recent report from MIT argues that “technological change is simultaneously replacing existing work and creating new work. It is not eliminating work altogether.”
In other words, automation has a transformative, rather than diminishing, effect on jobs. The authors of the report also offer a direct response to the question of job replacement by machines, based on an analysis of employment participation rates over time when compared to technological innovations:
The prospect of mass unemployment runs contrary to the evidence. Even as technological advances have made life longer, more comfortable, and more interesting, they have generally led to net job creation rather than net job destruction. The Future of Work: Building Better Jobs in an Age of Machines. MIT.Autor, Mindell, & Reynolds
Given this context, the real question about automation isn’t whether or not it will take our jobs, but what future role automation will play in our work.
Reality check: automation in the workplace
Dystopian scenarios in which robots and AI render the human workforce completely irrelevant aren’t supported by current data or historical trends. Still, the status of human-automation relationship status is likely to remain “it’s complicated” for the foreseeable future.
Automations do not appear poised to completely take over our jobs, but most of us should expect to see automations increasingly embedded into our workflows in the coming years.
Just how much of our work can be automated depends on the nature of the work. A 2015 analysis by Chui, Manyika, and Miremadi that only a small percentage of U.S. jobs (5%) can be fully automated. They also found that for about 60% of U.S. jobs, automation could save an average of 1.5 days of labor per week, or 30%.
Job potential for automation
A recent analysis by McKinsey considers which types of jobs have the highest potential for automation, as well as which are most insulated from its effects. They found that jobs involving “performing physical activities or operating machinery in a predictable environment” would see the highest degree of automation, in some cases as high as 78%. This should be unsurprising, given what we are already seeing in the manufacturing sector. For jobs that require unpredictable physical work, that percentage drops to 25%.
In other words, where critical thinking, problem-solving, or complex decision making is required, machines are no substitute for the human mind.
Other types of work in which automation is predicted to take on a larger role include data processing and data collection, two areas in which we are already seeing automation have an impact. Both data collection and processing depend on predictable, consistent workflows that can be replicated through automations.
Complementing, not competing
Physical labor, especially repetitive work that requires decision-making or creative inputs, will increasingly feel the shrinking effect of automation on the total amount of available labor. In other types of work, however, automation will play more of a complementary rather than competitive role. McKinsey found that jobs in which workers interact with customers, bring expertise, or manage others will see a smaller percentage of their time and activities automated compared to those performing jobs with higher degrees of predictability or repetition.
In these cases, automation reduces the amount of time workers spend on tasks that aren’t directly related to their work, such as sending emails, coordinating tasks and schedules, generating standard documents, recreating workflows, or manually updating deliverable statuses.
Even in jobs that require a high degree of specialized skills, automation can resolve tasks that are repetitive, predictable, or routine. The outcome is that people can contribute more holistically to their workflows and processes, instead of a constant, microscopic focus on their tasks.