As AI makes growing inroads into medical practice, many doctors worry that AI will take over key elements of their jobs.

In some industries, workers striking in part to limit the role of AI in their jobs.

Earlier this year, more than 15,000 nurses in the New York State Nurses Association struck for 41 days. They ended up with a contract that includes protections regarding the use of AI in patient care, as well as guarantees for pay raises, staff increases, and safeguards against workplace violence.

Mental health professionals at Kaiser Permanente in California also struck recently, in part over concerns their employer would replace them with AI, or that mandatory AI tools like ambient scribes would threaten patient privacy. The strike ended, but Kaiser Permanente has not agreed to the AI limits requested.

In the film industry, the Writers Guild of America held a 148-day strike in 2023 and won guarantees that AI would be not considered a “writer” and would not be used to write or rewrite “literary material.”

These modest successes raise the question: Should doctors consider labor actions to keep AI from replacing physician job skills?

Does AI Really Threaten Doctors’ Jobs?

A Medscape survey showed that 23% of doctors are apprehensive about the impact of AI on medicine. While most physicians expect AI to be an adjunct to medical care, others fear becoming redundant and seeing AI take over major elements of their job.

“I think most physicians worry about losing their jobs, not about losing some of the tasks they do every day,” said Bertalan Meskó, MD, PhD, director of the Medical Futurist Institute and author of The Guide to the Future of Medicine: Technology AND The Human Touch. “And even if the fear about losing jobs is real due to doomsday articles about the rise of AI, I see AI’s role in care more like an assistant we simply cannot afford to avoid.”

“Right now I don’t think there is a risk,” said Rebecca Andrews, MD, MS, CAHP, professor at University of Connecticut Health, Farmington, Connecticut, and Chair of the Board of Regents, American College of Physicians.

“When getting patient history, AI can’t match the physician at pulling on threads of stories and using repetition, experience, and volume of symptomatology to find the right diagnosis. When I’m in a room with a patient who gives an answer but looks away, I know there is more to the story, and AI cannot do that,” Andrews said.

While AI is useful in reading scans, human judgment and oversight is necessary, Andrews added. “There’s a risk of over-reading and over-diagnosing with too much technology unless there is human interpretation alongside of it.”

AI poses a bigger job threat to some specialties than others, said Gregory A. Hood, MD, internal medicine, Lexington, Kentucky. “Radiology, dermatology, pathology, and specialties that are image-based may face greater risk but not surgery or some other specialties,” he said.

Another danger related to growing dependence on AI is physician ‘de-skilling’, or losing some of their skills, said Hood.

“I know of residents who are under remedial action, or facing being fired, because they aren’t using their brains, aren’t learning or training,” Hood added. “They have been caught loading everything into AI and just doing whatever AI told them to do. These residents aren’t learning how to be doctors or learning the information that they need to know.”

Would Striking Over AI Even Work?

The concept of physicians’ strikes has always been controversial.

“I’m not opposed to the concept of striking at some level, but what are the clearly defined goals?” Hood said. “How effective is a strike going to be? A whole lot of doctors can’t afford to go on strike because they have patients, debts, high mortgages, and car payments. To be able to strike you’ve got to afford living during the strike,” said Hood.

Because AI is more of a threat to just a few specialties, the bulk of doctors are not likely to be willing to strike, said Hood. “A strike will give hospital administration an excuse to replace more doctors with physician assistants.”

Hood noted that resident doctors in the UK have struck repeatedly in recent years. “They’ve gotten some concessions, but they really lost the sympathy of the public. Once you start pulling those levers, you expend that goodwill capital pretty quickly and it doesn’t replenish easily.”

Meskó noted that AI’s wide implementation at hospitals is inevitable.

“Organizing strikes about something inevitable sounds like an attempt for a solution that arrives way after the problem appeared,” he said. “Instead, medical associations and healthcare governments should start educating the healthcare workforce about how to use this technology for their own benefits, so physicians can start spending their precious time again with patients, not typing data on keyboards. Replacing, or at least reducing, administration is clearly the lowest hanging fruit for AI.”

Alternatives to Labor Action

Knowing that the genie is out of the bottle, it’s best to find ways to use AI to improve workflow and patient care, physician experts say.

“See what AI is about. Learn how to use it like you once learned statistics and how to read a paper,” Andrews said. “Don’t let it take over clinical reasoning for you; it’s a muscle that needs to be utilized and we see the skill set grows with expertise over decades.”

Hood added: “In your specialty, having your knowledge base, your focus and connection to the patient, and the things that really matter, the more you make yourself indispensable, the more the AI is going to remain on the sideline where it can still be a helpful partner. That’s the more appropriate role.”