The Potential of AI in Revolutionizing Medical Care
There’s been plenty of excitement and concern about how the coming revolution in artificial intelligence could upend parts of our lives. One significant benefit could be changes in medical care. AI has been called medicine’s biggest moment since antibiotics, with the potential for diagnosing rare illnesses, developing new treatment options, and exciting drug discoveries. Brook Silva-Braga put these endless possibilities under the microscope.
Traditionally, new drugs have been discovered by creating compounds and then seeing if they fight disease, a painstaking, almost artisanal process. But now, that process is starting to look very different. Don Bergstrom, the president of research and development at Relay Therapeutics, suggests that a revolution in drug discovery could cure cancer in the next 50 years by bringing new approaches to once-impossible problems. The number of potential chemicals that could become drugs is larger than the number of stars in the universe. Traditionally, labs might test a collection of 100,000 different chemicals. Now, computationally, scientists can screen a trillion compounds overnight, selecting only a few hundred of the most promising for physical testing.
A key part of this selection process uses new computing methods to better understand the proteins they target. Relay Therapeutics, for example, has learned to differentiate between proteins causing disease and those that are healthy, minimizing side effects. Marsha Meron, part of a clinical trial for one of Relay’s drugs, is living proof. She has metastatic breast cancer and, thanks to this experimental treatment, experiences minimal side effects while enjoying an active life.
Within five years, a significant part of the drug development pipeline could be accelerated by AI, according to Harvard’s Isaac Kohane, a medical doctor with a PhD in computer science. Kohane believes that AI’s potential in medicine goes beyond drug discovery and could revolutionize patient interactions. For example, a mother in Michigan used ChatGPT-4 to diagnose her child’s rare condition after doctors struggled to identify it.
Despite these advancements, Kohane sees AI as both an endorsement of technology and a critique of the medical community’s limitations. Currently, healthcare may be “stumbling,” leading to misdiagnoses and medical errors. Kohane suggests that healthcare outsiders, such as computer scientists like Praveen Raj Bhukar, are pivotal for AI’s integration into medicine. Bhukar’s work has shown that computers can read X-rays as well as humans, and he’s working on Metaverse, an experimental system capable of diagnosing a range of conditions.
Bhukar believes this model could be ready for regulatory approval in two years, and while the impact on patient health and employment for doctors remains uncertain, he envisions a transformative role for AI in medicine. Kohane hopes AI will democratize specialist knowledge, making expertise accessible to all doctors—and even patients. However, he warns that if the medical community doesn’t take leadership in AI’s role, others, perhaps with motives other than patient care, may control its future. Kohane advocates for transparency, suggesting that governments should step in if AI companies aren’t forthcoming.
While the rise of AI may be complex and controversial, Kohane is confident that it will ultimately improve upon current practices. As he puts it, “AI’s use is going to be problematic but far better than what we currently have.”
For CBS Saturday Morning, Brook Silva-Braga, Cambridge, Massachusetts.
