ChatGPT passes the USMLE:
Not too long ago, the news was filled with stories of how ChatGPT passed the USMLE, or the United States Medical Licensing Exams, which are tests that all medical doctors must take and pass in order to practice medicine in the US. They are long exams and considered some of the hardest licensure exams individuals can take on a professional level.
There’s been a lot of chatter about ChatGPT in general, including the ethical, economical and philosophical ramifications of artificial intelligence. Some people have publicly called for slowing down our use of AI until we figure out some sort of regulation, as they fear the possible consequences are too dangerous. Other folks are simply excited about AI, like ChatGPT, and testing it in a variety of ways. For example, kids are using it to write essays for schools, though I’m not sure if teachers are pleased with that. Technical writers are using it to write grants and proposals and copywriters are using it to help craft better material, though I have heard that a few companies have already laid off writers after replacing them with ChatGPT. Researchers are using it to write abstracts and discussions, and anecdotally, I’ve been told it does a really good job. It even produces cheesy love songs and poems about anything we ask it to. (Though I will say it needs a little help in the creative department.) But what about AI and the more serious jobs, like practicing medicine? If ChatGPT passes the USMLE, what else can it potentially do in the world of medicine? Will it replace the need for human doctors? It’s hard to imagine such a world, but as they say….anything is possible.
In a recent episode of Causes or Cures (my podcast), I chat with Dr. Victor Tseng about his and his team’s study on how ChatGPT came extremely close to passing the USMLE. (That’s right… it didn’t actually pass it, but it came damn close.) The study was published in Plos Digital Health here.
In the podcast, Dr. Tseng tells us more about what ChatGPT is and its current capabilities, how they set up the study to test its performance on the USMLE, results of the study, future studies his company may perform, and perhaps most importantly, where ChatGPT could take medicine and public health in the future, while highlighting some of the major ethical and philosophical concerns. I ask the question everyone wants the answer to..will robots replace us?!?
Who is Dr. Tseng?
He is a practicing pulmonologist (lung doctor) and the Medical Director and Executive Vice President for AnsibleHealth, a digital theranostics and AI-enhanced medical practice focused on care for diseases of the lungs.
To listen to the ChatGPT passes the USMLE episode, please click here!
And thank you to everyone who has subscribed to the podcast and/or shared episodes with your friends. It means a lot to me. :)
Please also check out other episodes from the podcast, such as this one: