AI Could Become Medicine’s Right-Hand Man – The American Spectator

Patients have been turning to Google for diagnoses for years. Now, as artificial intelligence enters the medical world, their doctors may join them.

Hospitals and research centers have announced partnerships with AI creators like Google and Microsoft for the last several months. They seem intent on using AI to streamline secretarial processes, like the long letters hospitals write to insurance companies, but these organizations likely have bigger dreams for the future.

No-Show? AI Knows.

Last Tuesday, Duke Health and Microsoft announced that they are partnering to develop what they called “game-changing” AI health care tools to complete basic administrative tasks like optimizing clinic schedules and predicting no-show patients.

Eventually, the hospital would employ the new technology to summarize research, making it easier for doctors to keep up with rapid innovation in their field. AI would also be able to generate letters to insurance companies and answers to patients’ questions. (READ MORE: AI Comes for Fast Food)

“This is gonna sneak up on us and the world is going to be really different really soon,” Jeffrey Ferranti, Duke Health’s chief digital officer, said, as reported in North Carolina’s News & Observer. “We’re trying to get ahead of that curve and do it responsibly.”

Google is already well ahead of Duke and Microsoft. It allegedly began working with partners like the Mayo Clinic to test its medical chatbot technology Med-PaLM 2 in April. The program is already being used to summarize research, organize health data, and generate responses to medical questions.

Doctors have begun to use AI algorithms in hospitals to scan tests like cardiograms and X-rays to predict which patients are most at risk. In May, scientists discovered that they could use AI modeling to “decode” brain scans.

AI as a Medical and Research Assistant

While medical facilities are currently interested in using AI to reduce the time they spend on secretarial tasks, both the Big Tech companies creating these tools and their customers have expressed interest in turning AI into a medical assistant for doctors and researchers — and they are well on their way to doing so.

In May, Google published research on Med-PaLM 2, which indicated that physicians reviewing answers to more than 1,000 consumer medical questions typically preferred the system’s response to the responses of their colleagues in eight out of nine categories. Those same doctors, however, noted that the program also had a tendency to generate off-topic or false statements more frequently than the real doctors. (READ MORE from Aubrey Gulick: Will There Be a New Government Agency for AI?)

Just two months later, at the beginning of July, Med-PaLM 2 took the U.S. medical licensing exam and passed, an exercise Chat GPT-4 had attempted unsuccessfully in February. But while Big Tech companies and hospitals are making rapid strides toward utilizing the new technology in clinics around the world, experts say they aren’t ready to trust it quite yet.

“I don’t feel that this kind of technology is yet at a place where I would want it in my family’s healthcare journey,” Greg Corrado, a senior researcher at Google who worked on Med-PaLM 2, said, as reported in the Wall Street Journal.