Somewhere in the hospital, someone is the last generation to hold their job title.
When Will AI Take Over Jobs in Healthcare?
This is the most common question I get from journalists. It is an extremely complex question that provokes a lot of internal philosophical debates.
In previous writings I’ve published, I’ve tried to convey how complex the AI issue is, so when the topic comes up, I naturally want to ask many counterquestions.
First: are we talking about automation or AI? These are far from the same thing but can resemble each other. And what type of AI are you interested in — general or specialized?
Of course, I realize that most people are not looking for a lecture but a digestible answer that can become a compact quote. Unfortunately, I will never be able to deliver such a thing — as a researcher, developer, and caregiver, I see too many sides of the coin. I often end up debating with myself of what the correct answer is, and no version of me has won yet.
Most futurists base their predictions on the world we live in now and make an educated guess about how already known parameters will develop. It often turns out to be wrong. How many guessed ten years ago that authors and illustrators would be the ones to lose their jobs first? Most were very vocal that radiologists and administrative positions would disappear.
Fast forward ten years, and I have personally waited three months for test results (and will probably wait much longer) because there is a “shortage of radiologists,” according to the message in my inbox. Meanwhile, I know illustrators who have lost their income because their clients have switched to using generative AI that creates images with a click.
When making predictions about which professional group will be hardest hit by digitalization and automation, the harsh and unpredictable reality is not taken into account.
Soldiers on horseback with bayonets did not suspect they would become a memory in modern warfare. But somewhere at Sahlgrenska University Hospital (or any hospital for that matter) stands someone who is the last generation to hold their job title.
The questions we should be asking are:
Which jobs do we want to eliminate in healthcare because they lead to burnout?
Which jobs will be cut due to economic reasons, even though we want to keep them?
Which jobs will be irrelevant in healthcare as technology makes tasks faster and more patient-safe?
If we look at jobs that wear people out, it can include administrative tasks. Some may be important but resource-intensive. Others do not significantly contribute to healthcare improvement but exist to maintain a bureaucratic system. We clearly want to leave these tasks to machines. The same goes for heavy lifting and monotonous physical work.
Several hospitals in the US have acquired robotic assistants that deliver samples, restock supplies, and change bed linens. So far, they have been appreciated, but calling them AI would be a bit of an exaggeration. Hence my counterquestion: do you mean AI or automation?
If we look at cases where professions are cut solely due to economics, there are quite a few job categories that could be eliminated today, according to experts (if we completely ignored patient safety, that is).
But what is often missed by both the experts and the entities trying to automate away human jobs is that it costs a lot to develop, certify, and maintain these systems, not to mention all the preparatory work required to annotate data and develop new workflows around the solution. In today’s times of cutbacks, one might look at AI with longing eyes, but if you really understand how AI systems are built, you quickly realize that AI is an expensive affair.
Then there is the third question: which jobs will be irrelevant? Which tasks are better performed by machines? I think this is what most people actually wonder, and the answer is: all, and absolutely none. A really dull answer, I know, but that’s how it is.
To answer which tasks or professions will be irrelevant, we need to know which problem-solving abilities will be needed in ten, thirty, or fifty years, and that type of clairvoyance is better left to science fiction writers.
Those who haven’t lost their jobs to chatbot services, that is.
When will AI take over jobs in healthcare, you asked. The answer is: AI will take our jobs when it can solve problems that we ourselves do not want to, cannot, or should not solve.
Instead of asking when this will happen, maybe as a society, we should prepare for it to happen soon and seriously consider how we can do this as ethically as possible and with as much reverence for humanity as we can — whether we have answered the question or not.
