ACR Data Science Institute (DSI) Chief Medical Officer Bibb Allen Jr., MD, FACR, was highlighted in a Nature article about the artificial intelligence (AI) revolution in medicine.
Although many physicians are aware of AI tools, only 10-30% of them have actually used them, according to Nature. But excitement seems to be growing about “generalist medical AI,” which assesses every anomaly in a scan and assimilates it to something like a diagnosis, more in line with what a physician does, rather than serving specific functions like detecting lung nodules in a computed tomography (CT) chest scan.
“The real goal to me is for AI to help us do the things that humans aren’t very good at,” Dr. Allen told Nature.
Researchers are now exploring medical AI with broader capabilities, inspired by large language models like ChatGPT. These models, referred to as “foundation models,” are trained on diverse datasets and can be adapted for different tasks without the need for extensive annotations. They have the potential to address the limitations of first-generation medical AI tools, the article said.
Big tech companies like Google and Microsoft are investing in medical imaging foundation models that combine multiple types of data, like images and text. These models have shown promise in improving diagnostic accuracies and providing more comprehensive insights.
However, even the most advanced AI tools are not yet on par with human radiologists. While the diagnostic potential of AI devices is exciting, there is a high bar for success and it's essential to be responsible when integrating AI into clinical care, according to Nature.
Read the full article on Nature.