This post is contributed by Daniel Ortiz, MD, chief radiology resident at Eastern Virginia Medical School, vice chair of the ACR Resident and Fellow Section and member of the ACR RFS AI Advisory Group.
I recently received this email from a medical student (name withheld) interested in radiology, but concerned that artificial intelligence may negatively impact the specialty:
Concerns are circulating amongst my fellow medical students regarding AI replacing radiologists. These worries are deterring some of my classmates from applying to diagnostic radiology programs, out of concern for the future of the specialty. I am very interested in radiology and would love to hear what you think about how radiologists might utilize AI to improve the field. Do you have any resources that might help me dispel this myth among my classmates and reassure us about the future of radiology?
Here is the abridged version of my response:
It can be intimidating with some of the things controversial non-medical data scientists (and sadly some clinicians) are saying these days. The good news is that some of the hype is accurate.
Those possibilities are the part of AI that makes radiology one of the most exciting fields. Here is a blog written by a radiologist expert in machine learning (AI) that sheds light on many imaging AI myths.
Several aspects of what we do as radiologists are routine and mundane. These take away from the intense thought-provoking aspect of our jobs and contribute to burnout. They are the low hanging fruit to automate. For instance, I look forward to having automatic volumes of tumors that can be trended, rather than relying on manual caliper measurements.
It’s true. There are some tools being developed to help aid in image interpretation. However, the time and work necessary to make these concepts workflow realities is significant. This is the work the ACR Data Science Institute™ is spearheading to ensure the focus is on algorithms that are safe and help us take care of our patients.
When these tools are developed, they will be diagnostic aids to make our work more efficient. A great comparison is the “autopilot” tool for pilots. Autopilot has been around for decades and has made aviation safer, but the number of pilots needed has gone up (not down).
Such concerns are part of what led us to create the ACR RFS AI journal club. We make it a point to learn about the technology and what it can do for us. We are building a community of radiologists and data scientists to focus on learning about and developing tools to help us.
You are welcome to join future journal club webinar sessions. I also recommend you join our ACR community as a member-in-training. It’s the best way to get reliable, up-to-date information and get involved.
I would not worry about radiologists being replaced. If that day ever comes in the distant future, all jobs are at risk. I hope you follow your interests and join our exciting field and the ACR community.
I would be interested to know how you might have responded to this bright medical student.
Please share your thoughts in the comment section below or share them on Engage (login required).