A patient recently showed up to my emergency department after he’d been doing home repairs for several days, including lifting heavy equipment and moving boxes. When he started feeling a recurring tightness in his chest, he entered his symptoms into a large language model, which told him he likely strained a muscle. He rested and waited for the pain to go away, but it never did.
When he finally came to my emergency room a few days later, his EKG and lab tests showed that he had suffered a serious heart attack. Because of the long delay, part of his heart muscle was permanently damaged, and he now has chronic heart failure.
Because of the long delay, part of his heart muscle was permanently damaged, and he now has chronic heart failure.
At a Make America Healthy Again summit Tuesday that was co-sponsored by OpenAI and Anthropic, Health and Human Services Secretary Robert F. Kennedy Jr. boasted that artificial intelligence would soon give Americans a second opinion “much better informed than any doctor in the country.”
He’s wrong to say that, and my position doesn’t arise from any philosophical opposition to AI technology. To the contrary, not only am I an emergency room physician, but I am also U.S. chief medical officer at Keikku, a clinical AI company. I use AI in some capacity every shift. However, the patient who consulted a chatbot and suffered irreversible heart muscle damage that a cardiologist could have prevented is an example of why human expertise matters now — and will continue to matter for the foreseeable future.
The AI chatbot a person uses at home cannot conduct a physical examination, perform life-saving tests and, in my patient’s case, it didn’t know what an emergency medicine physician learns on the first day of training: New chest tightness after exertion is a heart attack until proven otherwise.
At the MAHA summit, Kennedy promised every American would soon have full access to their medical records. I fully support that goal. But then he claimed AI could “correct” public health advice about issues including Covid masks and vaccines and free the public from what he called “medical tyranny.” Vice President JD Vance warned against “doomers” and added that experts no longer have a “monopoly on knowledge.”
The implication is that human expertise is an obstacle and that AI is the solution. As a physician and an AI builder, I can say that idea is completely wrong.
I am not a “doomer,” to use Vance’s word. My role at my AI company is creating tools that remove administrative burdens in healthcare, help reduce physician burnout and augment them in their medical decision-making. AI also plays a key role in my emergency room today. Ambient scribes write up clinical notes so I can concentrate on the patient rather than on a computer screen. Summarization tools scan years of records in seconds, and algorithms highlight potential lung collapses on chest X-rays.
Radiology was one of the earliest specialties to integrate AI, with hundreds of algorithms being used to screen images. Years ago, developers famously predicted AI would make radiologists obsolete, yet demand for them is higher than ever. Software can highlight a suspicious shadow on an X-ray, but it still takes a clinician at the bedside to interpret that scan within the context of the human in front of them. While some politicians preach AI liberation from physicians, some of the actual use cases are far less noble. Insurers, for example, using algorithms to rapidly deny coverage for certain medical procedures.
The implication is that human expertise is an obstacle and that AI is the solution. As a physician and an AI builder, I can say that idea is completely wrong.
The irony in arguing that AI will free us from medical experts is that the consumer AI tools available to patients seeking a second opinion synthesize the research, guidelines and papers produced by the very experts Kennedy tells the public to ignore. AI cannot be an escape from human expertise because today’s tools are built entirely upon it.
To practice medicine means making difficult calls in the face of uncertainty. The profession demands that a clinician assess subtle physical signs, understand what a patient is not saying and take full responsibility when lives are at stake.
Kennedy and Vance promoting AI as superior to doctors is dangerous. When public figures lead patients to believe that a chatbot knows better than a physician, people decide to avoid necessary medical care and stay at home while time-sensitive conditions worsen. AI is making steady and meaningful progress in medicine, yet my emergency room still faces the same overcrowding and long waiting times it did five years ago. AI can sometimes be a very useful assistant, but it is not a doctor. By selling it as superior to human expertise, Kennedy and Vance will only end up harming the people they claim they want to see healthy again.
The post RFK Jr.’s dangerous disparagement of doctors’ expertise appeared first on MS NOW.
Author: Source Author
From MS Now.
Leave a Reply