As of this week, less than 45% of new podcast feeds are created by actual people. The rest? AI-generated. Many of these podcasts focus on health. One company managed to produce 20 wellness podcasts in just 75 minutes, covering topics like vaccinations and mental health, all narrated by fictional hosts.
There’s notable potential in using AI for health education. Researchers at Boston University developed PodGPT, a tool that utilizes over 3,700 hours of medical podcasts to answer health questions in clear, simple language. A pilot study showed that these AI-generated podcasts performed comparably to those made by humans when measured on educational metrics.
However, the rise of AI in healthcare isn’t without concerns. Recent research published in BMJ Open tested several popular AI chatbots on medical questions. Half of the responses were rated as problematic. While they handled straightforward questions well, open-ended inquiries often produced answers that raised red flags.
Carri Chan, a professor at Columbia University, estimates that AI health podcasts are about 90-95% accurate. But the inaccuracies can be critical. They might overlook important caveats, overstate risks, and rarely suggest when it’s best to consult a medical professional.
Interestingly, studies show that many people can’t differentiate between AI-generated and human-provided medical advice. In one instance, participants rated AI answers more highly than those given by doctors, even when the answers were flagged as problematic.
In a survey by the Annenberg Public Policy Center, 63% of Americans viewed AI-generated health information as reliable. Meanwhile, a Pew Research study found that uninsured individuals often rely on AI chatbots for health information, as they are accessible anytime and usually free.
But the lack of regulation around AI-generated health content is alarming. There’s no federal mandate requiring these podcasts to disclose that they are AI-created. Most platforms that host these shows don’t have robust systems in place to vet the medical claims made in them. Experts suggest strict review processes and clear disclosures for AI podcasts, but these practices are far from widespread.
When listening to health podcasts, look for the host’s credentials. If you can’t verify who they are or if they’re a fictional character, be cautious. A reliable podcast should acknowledge uncertainties and when to consult a doctor. If every statement sounds overly confident, that’s a red flag—it’s a common trait of AI communication.
AI-generated health content may improve over time, offering valuable insights into patient education and making medical research more accessible. But right now, the rapid production of these podcasts outpaces quality control and regulation.
The transformation of health advice delivery through AI is significant, and many listeners remain unaware of the implications. Always cross-check medical information with a qualified professional before taking action.
