Editor's note: This column has been updated to provide attribution to original sources.
A McKinsey report from Michael Chui, James Manyika, and Mehdi Miremadi looked at the detailed work activities of more than 800 occupations across 54 countries to study the technical potential for automation and found that almost half of work activities globally could be automated using current technology. However, just because something can be automated doesn't mean it will be. The cost of developing and deploying the hardware and software for automation must be considered, as well as the supply-and-demand dynamics of labor. Other factors include the relative superiority of automation, and regulatory and social acceptance of automation for any particular task or occupation.
Jobs that are most vulnerable to automation are highly structured tasks in predictable environments. More than three-quarters of such activities could be done with today's technology, mostly in manufacturing and food service. Robots are able to do product assembly and packaging better and more cheaply than humans, and in food service, restaurants offer self-service kiosks or robot servers. Even though industrial robots are replacing humans in some routine manufacturing tasks, other low-skilled jobs (such as operating cranes or making hotel beds) are still better performed by humans and are at a lower risk for automation.
The report also found that the hourly rate wage rate alone is not a strong predictor of automation, despite some correlation between the two. Chief executives have a 25 percent automation potential while landscaping and groundskeeping workers have only a 10 percent automation potential. The reason that landscapers are less likely to be replaced by robots than other high-physical-labor occupations is that landscaping (and other activities that require physical movement in unpredictable settings) are more challenging to automate.
Also, analyzing work activities rather than occupations is a more accurate predictor of automation. "Overall, we find that only about 5 percent of occupations could be fully automated by adapting current technology," write the authors. "However, today's technologies could automate 45 percent of the activities people are paid to perform across all occupations." Humans are still better than robots at: spotting new patterns, logical reasoning, creativity, coordination between multiple agents, natural language understanding, identifying social and emotional states, responding to social and emotional states, displaying social and emotional states, and moving around diverse environments.
The hardest activities to automate are those that involve complex interactions with people, such as managing and developing employees, or communicating with customers, suppliers, and other stakeholders. Other jobs that are less likely to be replaced by robots are those that require creativity or expertise, such as artists and educators. "As technology races ahead," according to a report from Oxford's Martin School, "low-skill workers will reallocate to tasks that are non-susceptible to computerization--i.e., tasks that required creative and social intelligence."