Estimates of job loss due to automation miss one big factor

Predicting the scale of AI-induced job losses has become a cottage industry for economists and consulting firms the world over. Depending on which model one uses, estimates range from terrifying to totally not a problem. Since 2013, the most cited predictions include a study from Oxford University that predicted 47% of US jobs could be automated within the next decade or two; an Organization for Economic Cooperation and Development (OECD) report that suggested just 9% of jobs in the United States were at high risk of automation; PwC research that found 38% of jobs in the United States were at high risk of automation by the early 2030s; and a McKinsey report that found around 50% of work tasks around the world are already automatable. A lot of this difference can be explained by the methodology of these studies. Looking at whether an entire occupation can be automated is a different question than whether tasks within that occupation can be automated.