Smith also raises the issue of subjective assessment in older studies. And this seems that determine the risk of job replacement.
He cites a study by Frey and Osborne in 2013, where researchers hand-labeled jobs from a database developed by the US Department of Labor as either “automatable” or not.
Smith claims subjective bias can affect such assessments. And he adds that studies have since improved their methodology.
Next wave automation
Smith references a recent study by Goldman Sachs that takes a more holistic approach to assessing AI’s impact on automation.
Instead of focusing on entire jobs, the study views jobs as a sum of tasks described in a government database. The study acknowledges that automation can complement instead of replacing a worker’s efforts. It works best when only certain tasks are automated.
Best case, technology can create new tasks and jobs, which makes workers more productive in performing other tasks.
However, Smith points out that even with such findings, some media outlets still sensationalize the results with headlines that perpetuate the “robots take our jobs” narrative, disregarding the nuances of the study.