Skip to main content

AI and robots will destroy fewer jobs than previously feared, says new OECD report

AI and robots will destroy fewer jobs than previously feared, says new OECD report

/

But the impact will still be significant, increasing societal division between the rich and the poor

Share this story

Funding For Scotland's Decommissioning Fund Announced
New research says we may have over-estimated how easy it is to automate jobs involving manual labor.
Photo by Robert Perry / Getty Images

Fewer jobs are at risk of automation from AI and robotics than previous forecasts have warned, according to a report from the OECD, an inter-governmental group of high-income countries.

The new study offers a counterpoint to an influential 2013 paper by Oxford University academics Carl Frey and Michael Osborne, who warned that around 47 percent of jobs in the US were at high risk of being automated. Frey and Osborne’s research set the tone for much of the recent debate over automation, and its message has been reiterated in subsequent studies.

But, according to the OECD’s analysis, these current fears are somewhat overblown. The researchers found that only 14 percent of jobs in OECD countries — which includes the US, UK, Canada, and Japan — are “highly automatable,” meaning their probability of automation is 70 percent or higher. This forecast is far less dire than Frey and Osborne’s, although it is still significant, equating to around 66 million job losses.

In America alone, for example, the report suggests that 13 million jobs will be destroyed because of automation. “As job losses are unlikely to be distributed equally across the country, this would amount to several times the disruption in local economies caused by the 1950s decline of the car industry in Detroit where changes in technology and increased automation, among other factors, caused massive job losses,” the researchers write.

Inside Turkey's First Humanoid Robotics Factory
Despite hyped coverage, robots and AI still can’t handle tasks involving social skills and creativity.
Photo by Chris McGrath / Getty Images

But why is this new estimate so different from Frey and Osborne’s? One of the reasons is the OECD’s attention to variation between jobs of the same name and to tasks that are difficult for computers to manage even within highly automatable jobs.

Think about a machine operator in a factory, for example. Although a portion of their job could be automated, they may have other responsibilities (such as managing inventory and overseeing junior workers), which computers cannot manage. And consider the difference between a worker in US garment factory and its equivalent in Vietnam: the American factory is more likely to be technologically advanced, and the typical worker’s day will likely involve a greater number of non-routine tasks that resist automation.

All of this is to say that predicting how and where automation will have the greatest impact is an extremely difficult question. It’s perhaps part of the reason that Americans consistently say that automation will take other people’s jobs and not their own: we all have a much clearer picture of the variety of skills our own jobs’ demand.

But although the OECD predicts that total job losses will less than feared, the report stresses the impact will still be incredibly damaging to groups already under threat in today’s labor markets: low-skilled workers and the young.

As the researchers write: “The risk of automation is not distributed equally among workers [...] Occupations with the highest estimated automatability typically only require basic to low level of education.” This also means the effects of automation are felt disproportionately by the young. The researchers note that although we assume young people are better placed to adapt to new technologies, almost 20 percent of people aged 20 and below in OECD countries work in low-skilled jobs, like cleaning and food preparation, while 34 percent are in sales and personal services — all jobs likely to suffer from automation.

These trends are particularly important as they show how this coming wave of job losses could further polarize society between high-paying, high-skilled jobs, and low-paying, insecure occupations.

Automation will increase social inequality and hurt the poor the most

Previous research has come to similar conclusions, with a study published last July in the UK showing that AI and robotics are likely to stifle social mobility. This is because of a number of overlapping factors, including the fact that those in low-paying jobs are less likely to be able to retrain for new professions and because automation is likely to kill off many routine jobs that were stepping stones into professional careers for people with low education, like doing admin work in a legal office, for example.

The OECD’s research, which was published last month but resurfaced by a recent story in the Financial Times, stresses that “technology will without doubt also bring about many new jobs,” but in light of these forecasts, governments need to focus on providing money and resources for education and training for the groups most at risk.

“Re-qualification is an important mechanism to aid the transition from more to less automatable jobs,” write the researchers, adding that it is “important not to dismiss the importance of providing retraining and social protection” for young workers and those in low-skilled jobs.

The coming wave of automation may not be as destructive as we have previously feared, but that does not mean we can afford to be complacent.