The Trump Administration is moving ahead with plans to integrate medical artificial intelligence (AI) into the U.S. health care system. Officials are greenlighting efforts to allow AI agents to diagnose and prescribe medications and therapies to patients. However, some stakeholders and veteran HHS regulators worry that top officials are allowing the tech industry too much latitude in rolling out AI medical providers, despite no experience in federal health policy. According to The New York Times:
Chris Klomp, who faces a confirmation hearing this week to be the second-in-command at the health department, said on a podcast in May that the use of A.I. in patient care was the “holy grail” in the agency’s quest to improve the nation’s health and reduce costs.
One Silicon Valley venture capitalist believes AI can replace doctors, telling conference attendees:
“It’s over for doctors, human doctors, A.I. is just going to be better.”
He added, “C.M.S. seems to be very excited,” that an A.I. primary care provider could give underserved people rapid access to care.
Skeptics – especially doctors at the American Medical Association (AMA) – counter that AI is unproven and “are not ready for prime time.” They argue that AI can never replace human doctors. Yet doctors often turn to AI for diagnostic second opinions, for administrative tasks, medical records charting and patient encounter note taking. The CEO of the AMA has argued:
“A.I. in health should not be driven by the tech industry,” Dr. Whyte said. “It should be driven by the clinical community. And right now, what we have going on is that tech is the tail wagging the dog.”
While Dr. Whyte has a point, it could also be argued that the profession most impacted by medical AI should not be allowed to impede it. Participation in testing medical AI by medical professionals is necessary, but it should not be ignored that both tech company executives and medical professionals have distinct biases. For their part, tech company vendors are trying to prove the quality of their products:
Curai Health, was selected by Medicare and has an application pending with the F.D.A. to deploy its autonomous A.I. primary care doctors. In a pre-published study by company scientists and engineers, his team found that clinicians concurred with the AI doctor’s diagnosis about 91 percent of the time across about 2,300 cases. However, in three cases, the A.I. suggested a telehealth visit in situations where a human doctor advised an urgent or emergency evaluation.
The deputy commissioner for technology and artificial intelligence, a political appointee with ties to the tech industry, clashed with agency officials who wanted a more measured approach within the FDA. One of those agency officials was the director of medical devices, the department that oversees software that involves patient care. Many medical devices contain software. Just think of any household electronic device. They all contain lines of code to control how the device works. It is the same with medical devices. The software must be tested and validated. It is hard to imagine software with more potential for harm than an unvalidated artificial intelligence provider directing patient care. Yet, the tech industry wants less oversight into AI. Industry proponents of medical AI even tried to get the director of medical devices fired.
It remains to be seen how medical AI will be funded once integrated into the health care system. Secretary Kennedy’s son, Finn Kennedy, is a venture capitalist who works in med tech AI. He has stated that AI doctors won’t work for free. But then how will they get paid? How much will patients be charged? Will medical AI only become available at hospitals charging hospital prices and program AI to practice medicine ways that benefits hospitals at the expense of patients? If so, the consequences will be dire. Physicians earn significant salaries because mastering medicine is difficult, medical education takes many years to learn, there is a physician shortage, and it is impossible to quickly increase the supply. By contrast, AI is scalable. Or perhaps AI office visits will become a subscription service or based on data usage charges like is common today.
It is easy to assume that medical AI will be only used as a primary care chatbot, with prescriptions approved by a physician. That is just the beginning. Medical AI is worth pursuing but it should benefit patients, not just those who own the systems.
Read more at NYT: Trump Administration Moves to Integrate A.I. Into Medical Care Despite Concerns