Challenges of the changing work through AI and robotics
No, a robot will not take over your job
Yes, although some of the tasks in many current middle-skilled jobs are susceptible to automation, middle-skilled jobs will continue to demand many different tasks of all sorts.
Hard to Automate Tasks
Abstract and Manual Tasks
Abstract tasks include those which require problem-solving capabilities, intuition, creativity, and persuasion. As well as manual tasks that would normally require situational adaptability, visual and language recognition, and in-person interactions. These are not ‘high-skilled’ for the U.S. labor market standards, but it is very challenging to automate them. From the abstract workers category, those which have task-intensive occupations that complement between routine and abstract tasks are to benefit from information technology.
Adaptation rather than substitution
But let’s stay wary…
Although automation can lead to adaptation rather than substitution in workers, one has to be careful of the transitional process. For example, when a machine replaces a job that a human was conducting, leaving them temporarily unemployed, there is a necessary transition in between. If said person has to get skilled in a new area in order to ‘adapt’ to the new working world, who will provide a salary for them in the meantime? How hard will it be to find a new job, being that they are not the only unemployed person? As the price of computing power has fallen, computers and robots have increasingly displaced workers in accomplishing codifiable tasks (‘routine tasks’).
This has led to a big decline in administrative employment, and, to a lesser degree, in production and operative performances. It is true that the interaction between technology and employment requires thinking about more than just substitution, but people that have worked their whole life in one sector will find a hard time relearning and finding a new suitable job, even when theoretically they can, as automation can/should/will bring a whole other sphere of job opportunities.
Another interesting and not so obvious implication of automated systems is that, even with technology being so involved in the process, these are intrinsically man-operated systems, and both human and technical factors are on the table. A concern that this brings is that of control in the process industries, for example in monitoring, where a high-skilled employee has to guarantee safe automated operations and be present in case of need, but does not have to actively exercise any of their skills, so over time these are lost or a decrease is shown. Humans also tend to be in better health when working low peace, and these control jobs usually work with fast process dynamics and high frequency of actions. High levels of stress lead to errors, poor health and job dissatisfaction.
Takeaways and recommendations to policy makers:
Job training systems should encourage the kind of skill workers can thrive with in jobs from the technological era. Human capital investments have to be at heart of any long-term strategy, and technological changes should aim at complementing rather than substituting human workers.
Distribution issues is a very likely upcoming challenge. Some researches explore multigenerational economic environments in which some few generations can enrich themselves with the burst of robotic productivity, leaving after-damage in future generations. A recommendation to give to policy makers is to create a capital tax that makes the technological advance progressive and welfare oriented.
A big challenge occurs in hiring processes concerning the ambiguity by which algorithms read and perceive anti discriminatory laws and regulations. Some interpretations of the disparate impact doctrine (laws that, even when neutral, adversely affect one group of people over another) are ill-suited to address biases that arise in machine learning models. For that, the state has to provide more transparency in the process and stop overseeing and assessing discriminatory hiring tools, even when unintended.
And if you want further lecture, I recommend the following bibliography as a departing point:
- Autor, David H. n.d. “Why Are There Still So Many Jobs? The History and Future of Workplace Automation.” Journal of Economic Perspectives 29 (3): 3–30.
- Bainbridge, Lissane. 1983. “Ironies of Automation.” Automatica 19 (6): 775–779.
- Boegen, Miranda, and Aaron Rieke. 2018. Help Wanted: An Examination of Hiring Algorithms, Equity and Bias. Washington, D.C.: Upturn.
Thanks for reading! It is my first time writing here :)
