My 'rethinking medical education' post was written 100% by ChatGPT
It was a lot like me, but not me.
I am actually writing this column. But I did not write a column that appeared last Friday, entitled Rethinking Medical Education. That column was 100% written by Chat GPT.
It all got started with Timothee Olivier. He had a wild idea. He asked Chat GPT to write a column in my style. (Below is his verbatim prompt). He suggested I post it to see the response. He made no edits to it. The column I posted was that.
When I read it the first time, it sounded a lot like something I would say. Because I said something similar 10 years ago. That article was published in academic medicine. Of course, Timothee knows my views on this topic, so he gave some guidance. At the same time, the essay that Chat GPT wrote is different than what I would write. It is drastically different than my 2012 essay.
The essay generated lots of comments, and no one thought that I didn't write it.
What's the lesson: software like chat GPT and Bard have made remarkable progress. At this pace, they will soon become indistinguishable from the people they emulate. Look how close we are already.
What does this mean for creative writers? I continue to believe that the one thing that they will not be able to do is generate the original kernel of the idea. On this topic, that kernel was generated more than 10 years ago. Yet Chat GPT fleshed it out nicely.
Timothee messaged me a day after the post went out expressing regrets. Some people may feel betrayed or fooled. That was not his or my purpose. And I understand that feeling. When you read someone you want to feel like they are with you-- not a simulacrum of them.
I will be writing my columns in the future just as I have done in the past, but I agree with Timothee that this was a useful lesson to show the power of the technology.
What do you think? (Leave a comment)
Here is the prompt:
Hello, ChatGPT, could you write a 1000 words essay, in the style of Vinay Prasad, MD- MPH, writing in the style of his personal substack, punchy as usual. The whole idea of the essay would be that medical training should be shorter, but focused on evidence-based medicine and patient's care very early on. One paragraph would be about the fact that because biology made great advances during the 20th century, it became central in medical education. The second paragraph should emphasize that the other main pillar of medicine that occured in the 20th century is evidence-based medicine, with randomized clinical trials at its core. Third, we should underline that timing in the curriculum matters : if students are first tought about the primacy of biology, they will later reject evidence-based result based on implausibly biological findings, yet what matters to patients is what is beneficial to them, whatever the biological reason. In a last paragraph, Vinay would argue that the medical curriculum should start with history of medicine, great principles of trials, examples of medical reversal, when biology was later reverse in randomized clinical trials like in CAST, and that medical eduction should stand first on evidence-based principle before teaching basic principles of biology and anatomy. Make more paragraphs than less if necessary. Avoid the term "paradigm shift". The style has to be easily understandable for lay people also. The first sentences should be short and clear about the whole argument of the essay.
You need to include a typo in each column just as proof of life ; )
I was pretty unimpressed with that post, tbh. Don’t agree with some of it (grasp of biology has to come first) and much of it was boring. The real Vinay does better.