If you’re asking if AI will replace radiologists, you’re asking the wrong question

There are two ways to make certain that you will get people to click on your article in the radiology world right now – either write that AI is going to replace radiologists, or argue that AI will absolutely, certainly not replace radiologists.

It’s easy to see why either of these positions are so eagerly consumed. On the one hand you’ve got robots replacing humans, and not just for repetitive, physical tasks, but even for cerebral activities like being a doctor. No one is safe! And whenever one take on a topic becomes popular, the opposite position also gains interest – humans fight back against the robotic takeover! Either extreme of the argument will get readers.

We need to have much more nuanced conversation, one that goes beyond the headlines. Others like Luke Oakden-Rayner have done a great job in doing just that, going deep into the topic. We should be looking at task replacement, not job replacement. And if we care about the overall radiologist job market, we should be thinking more deeply about all the factors that go into that. I was also encouraged by the recent SIIM MIMI (Society for Imaging Informatics in Medicine – Machine Intelligence in Medical Imaging) meeting, where the question was disposed of relatively quickly, although a hint could still be found in the presentations and conversations.

Many who have been thinking about this topic have realized that it’s the wrong question. But once we’ve decided that it’s not the right question, with what should we replace it?

First of all, why do you care? If you really are just interested in 20 seconds of controversy as entertainment, it’s probably fine as is. But chances are, you are in a particular situation which makes it more important to you.

Here are a few situation-specific alternative questions you might ask yourself, according to your role.

Current Radiologists:

Are you a current radiologist, worried about your future? What are you really worried about – robots taking your job, or job security in general? While AI may reshape the nature of the radiology job market in general, your particular job probably has other, more pressing risks associated with it. You might instead ask yourself:

How can I add the most value?

Chances are, you are in much more immediate danger of being replaced by other humans, similar to yourself. Other humans who can add more value than you are – faster, better, or cheaper. If you are always thinking about how you can increase your value proposition and then doing something about it, you will be much more likely to keep your job despite any other competitors – robotic or otherwise.

Putting AI aside, are you using all the tools currently available to provider the most value? Are your reports easy to understand by ordering docs and easy to use for comparison by subsequent radiologists? How do you know? How do you help others know what exams to order? Do you help improve the quality of your exams? If you are engaged with not only doing the work but making it better, AI will simply be another mechanism for doing so.

Potential radiologists:

Are you a potential radiologist, wondering if there will be a future for this career? It’s very hard to predict what any job market will do, especially when considering a technology in its infancy, like AI in radiology. It will almost certainly replace some tasks that radiologists perform, many of which would be quite welcome. That may affect the job market for radiologists. But whether the overall job market will shrink or grow in an absolute sense is hard to predict. And remember that market forces, though potentially concerning, probably have less to do with your personal future than your individual situation. So instead of wondering if there will be radiologists in the future, you might ask instead:

What kind of career do I want?

Radiology has always been a rapidly changing field, much more so than many other areas of medicine. It changes at the speed of technology, not the speed of the development and approval of new drugs. Although there are many radiologists who are resistant to change, in my experience the most successful ones are those who adapt and adopt changes into their work.

If you are already thinking about entering into radiology because you like the nature of the work, consider if you are OK with frequent change and finding ways to adapt that change to your workflow. Are you curious as to what the next new thing will be, and can’t wait to figure out how to use it? Or do you want to be doing the same work in retirement that you are doing right now? If the idea of the radiology roller coaster is exciting, then welcome, and strap in.

Those working in technology:

Are you in tech, looking for a new possibility? Maybe you are already in the radiology space, and have products that could be enhanced. Maybe you want to enter into this new field and wonder where you can get involved. The question you should not be asking yourself is how you can replace the people in healthcare. Not only because it tends to make your customers defensive, but because it’s also not a great approach. Presumably you want to create useful tools, help improve the work, and make some money while you do it. Replacing people per se may or may not occur, but it’s not really the goal, is it? Why not ask instead:

Where can I add value into the system?

AI brings a lot of potential to improve the process of healthcare. The job of any healthcare provider – physician, nurse, technologist, MA, scheduler, anyone – is filled with nonclinical tasks that could better done by a computer. Pick any process in medicine, and you will find people performing tasks that can be performed much faster, more accurately, and more reliably by computers than by people. And for many of these tasks, people want you to do it for them. Does anyone complain that they’d rather scan an entire page themselves instead of using a “find” feature to locate what they want in the page? Does anyone wish they had to perform long division by hand, cursing the burden of calculators embedded in every computer and phone? If you want to understand where you can add value, you’re going to have to gain an understanding of the work as it’s currently done.

You may have noticed that the questions for tech folks is similar to that for radiologists. It’s a theme. The question may be the same, but your answers will vary depending on your specifics. Finding where you can add value is always a good idea.

While people are busy arguing at the relatively shallow level of whether or not radiologists will be replaced, others are looking beyond and asking questions that matter, questions that take some thought to both ask and answer. How deep do you want to go?

2 thoughts on “If you’re asking if AI will replace radiologists, you’re asking the wrong question

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s