This browser does not support the Video element.
A striking new study from Mass General Brigham revealed that ChatGPT is nearly 72 percent accurate in making medical decisions, including identifying possible diagnoses and final care decisions.
OPENAI, CREATOR OF CHATGPT, ON TRAJECTORY TO BRING IN $1 BILLION
Some critics argue that artificial intelligence studies aren't grounded in real clinical needs, but not Dr. Marc Siegel. The New York University professor of medicine joined "The Big Money Show" to bolster his support for A.I., while also warning doctors to "remain in control" of the powerful technology.
"I actually like it, and I think overall the balance is in favor of it. And I'll tell you why," Siegel teased. "This study is from the Journal of Internet Medical Research, and they look at the Merck manual and 36 scenarios from Merck, and they look at it from beginning to end. What could this be? What is it? What do we do? What's the final diagnosis in there? The more complicated it was, the less well ChatGPT did. It's somewhere between 60 percent and 72 percent."
AMAZON TOUTS ARTIFICIAL INTELLIGENCE PRODUCTS, SERVICES FOR BUSINESSES
The study was conducted by asking the chatbot to determine a possible diagnosis for 36 standardized "clinical vignettes." The A.I. bot was able to determine a medical diagnosis accurately about 72 percent of the time, based on the patient's symptoms and personal medical information. Researchers found that ChatGPT performed best in making final diagnosis, where it was 77 percent accurate.
On the other hand, researchers were also able to pinpoint ChatGPT's greatest weakness – making deferential diagnoses. In this category, the chatbot was only 60 percent accurate. When it came to making clinical management decisions, such as figuring out what medications to treat a patient with after the correct diagnosis, it was only about 68 percent accurate.
FILE - An image shows screens displaying the logos of OpenAI and ChatGPT. (LIONEL BONAVENTURE/AFP via Getty Images)
Dr. Marc Siegel went on continuing to express his support for artificial intelligence.
"I think it's a clinical tool. Right now it's a research tool, it can help me with radiology, with cancer diagnosis. But in the doctor's office, especially if there's a scarcity of people for me to refer to specialists. Let's say I'm in a rural area and I can't get a diabetes doctor on speed dial or an oncologist on speed dial, this may help provide more information with decision-making. So, I look at it as positive, but doctors have to remain in control of it. It is not to replace us," he stressed.
AI-POWERED INVESTMENT PLATFORM BECOMES FIRST NON-HUMAN FINANCIAL ADVISER REGULATED BY SEC
A massively impactful finding from the study was the chatbot's ability to increase efficiency in healthcare – a critical necessity for medical professionals.
"Healthcare is 66 percent of the GDP, and it keeps going up every year. And now, more and more burgeoning Medicaid prices. And the government is paying a lot. And I think that on the surface, it would appear to improve efficiency – a lot – and save money," Dr. Siegel said earlier this week.
Although the capabilities of artificial intelligence seem infinite, Dr. Siegel pointed out that a large part of his job is about "compassion," something that ChatGPT could never emulate.
"The problem is that a lot of what doctors do is very nuanced, isn't it? It's about compassion. It's about what I called the ‘by the way’ diagnosis. It sets up one way, but then the patient gets comfortable with you and reveals something they weren't intended to reveal. And that ends up being their whole future right there. You can't replace that with a computer. A computer is not bringing empathy or creativity to the situation. So it's about efficiency," Dr. Siegel concluded.
GET FOX BUSINESS ON THE GO BY CLICKING HERE