AI may take over doctors’ tasks sooner than thought: Gottlieb 

Former Food and Drug Administration (FDA) Commissioner Scott Gottlieb said artificial intelligence (AI) could take over tasks from doctors sooner rather than later.  

“The inevitable question isn’t so much if but when these artificial intelligence devices can step into the shoes of doctors. For some tasks, this medical future is sooner than we think,” Gottlieb wrote in an opinion piece published with CNBC.  

Gottlieb said AI tools in health care are split into two categories: machine learning, which utilizes algorithms to allow systems to “learn patterns from data and make predictions;” and natural language processing, which understands and creates human language.  

In a few cases, Gottlieb said large language models are analyzing a patient’s medical records and providing diagnoses and treatments directly to the patient without a physician involved.  

“The biggest hurdle may well be establishing a suitable regulatory path,” Gottlieb wrote. “Regulators are hesitant, fearing that the models are prone to errors and that the clinical data sets on which they’re trained contain wrong decisions, leading to AI models to replicate these medical mistakes.”  

Gottlieb said overcoming these hurdles “holds significant promise,” to improve patient outcomes and tackle financial challenges that incur when more non-physician providers are hired to lower the costs of provider labor.  

Gottlieb said large language models are also being used for administrative tasks like processing medical claims, analyzing medical records or for clinical decision support software, which takes patient’s specific data and recommends diagnoses and treatments. Gottlieb said if the doctor is involved, the FDA may not regulate the tool.  

According to Gottlieb, machine learning technology is used to analyze clinical data, images and scans, and the software of which is often categorized as medical devices by the FDA. Gottlieb noted the tools are trained with closed data sets where the findings have been verified, which “gives the FDA increased confidence when assessing these devices’ integrity.” 

“Even if the current stage of development isn’t quite ready to completely remove doctors from the decision-making loop, these tools will increasingly enhance the productivity of providers and, in many cases, begin to substitute for them,” Gottlieb said.  

Gottlieb also pointed to how OpenAI’s ChatGPT, a large language model, passed the U.S. Medical Licensing Exam, while 10 percent of medical students who take it do not pass every year. 

Gottlieb’s argument comes as more studies reveal AI could take jobs from people in the future.

Last week, a report from McKinsey Global Institute showed that nearly 30 percent of hours currently worked across the U.S. could be automated by 2030.