AI may also replace skilled professions

Woman with helmet working in industrial setting.

Mainly simple and monotonous tasks in industry have been taken over by robots, but with AI also skilled professions may be automated. Photo: Getty Images

When work tasks are replaced by new technology, changes in the labour market follow. Research shows that this has historically led to more jobs overall rather than fewer. Mainly simple and monotonous tasks have been automated, but with the entry of AI into the workplace, this situation may change.

“Looking to the future, we can really see that this will affect several professional categories that require a lot of education, such as programming and certain types of doctors. After all, AI exists that can analyse X-ray images and reduce demand for radiologists. AI is also very good at reading and analysing huge amounts of text and could replace some lawyers. ChatGPT can make reporting easier, which will affect journalists,” explains Sofia Hernnäs, researcher in Economics at the Institute for Housing and Urban Research.

Robots have taken over

Her research is focused precisely on how automation has affected the labour market, and she has predominantly studied the period 1985–2016. Among other aspects, she has examined the outcomes for people who worked in professions where technological solutions completely or partially replaced humans, for example in industry where robots have taken over more and more of the heavy, monotonous jobs, or in offices where typists have become redundant due to computerisation. Their salary growth was affected.

“Among those whose occupation was affected, we found that the average employee lost approximately two to five percent of their income over the following 30 years. Our interpretation is that the average income earner is quite good at adapting. They change jobs a little more often and receive slightly lower total lifetime incomes, but not by much,” continues Hernnäs.

Porträtt Sofia Härnäs

Sofia Hernnäs, researcher in Economics at the Institute for Housing and Urban Research.

However, she noted that things were worse for those who had the lowest salaries in the declining occupations. On average, they lost around one tenth of their expected lifetime income. And the effects of being in a declining profession can have far more serious consequences than this; according to a new study on which Hernnäs is working, it could even shorten lifespans.

“Those who worked in reduced occupations had on average a five to eleven-percent higher risk of dying in the next 30 years, and I see this difference in those who are the lowest paid in each occupation here too. They are at an even higher risk of dying, which means premature death because these people are only 64 at most when we look at them,” notes Hernnäs.

High-status occupations may be replaced

In the future, technological solutions may well even place those in qualified, well-paid, high-status occupations – people who are often highly specialised in their fields and have honed certain skills.

“It may be that they find it more difficult to change professions. They would perhaps lose more relative to their basic position. Say you are a radiologist who earns SEK 75,000 a month and have invested ten years of your life in becoming a radiologist; you are losing a lot. That said, it’s not certain that these people will become poor compared to someone who has a lower paid profession. But those who have invested a lot will lose their status,” says Hernnäs.

Not fewer jobs so far

On the other hand, she notes, automation in the labour market has historically not resulted in fewer jobs overall. This may very well be the case even now as AI takes over more and more tasks.

“Certain technological developments remove work tasks, but other developments also create tasks where humans may have comparative advantages compared to machines, i.e. things that are more difficult to automate. This could, for example, involve managing these machines or the AI itself. I also imagine that care professions and the type of jobs that cannot be automated will increase in scope when we become richer as an economy and have greater demand for these types of jobs,” concludes Hernnäs.

Åsa Malmberg

Subscribe to the Uppsala University newsletter