top of page
  • Writer's pictureSociety of Bioethics and Medicine

Disserations, Virtual Machines, and now Medicine? ChatGPT Does It All.

Written by Jonathan Gao

Edited by Anling Chen

While artificial intelligence (AI) isn’t exactly new, the release of ChatGPT, an AI software released by the firm OpenAI in December, seems to have taken the entire world by storm - reaching 100 million monthly users in January 2023. The software is capable of completing tasks in response to a variety of user inputs. Think virtually anything from “Can you write me instructions on how to design a website like I’m five years old?” to“Write me a jazz song about Nordic vikings going to war with American soldiers.”

By simultaneously digesting large swaths of information found online and learning on-the-go from responses it gives to and gets from users, ChatGPT is becoming the go-to tool for many who are looking for general assistance, work revisions, or thought-provoking ideas.

Streamlining the most arduous, tedious parts of work is one of ChatGPT’s many key facets. Joe Bromley, a fashion journalism student based in London, used this to his advantage while writing his dissertation. Rather than spend hours combing over the Internet for a handful of sources, Joe asked ChatGPT to produce a comprehensive list of books and papers that he can read and cite - by doing so, he can direct his efforts towards thoroughly crafting his thesis.

On top of assisting with thesis papers, entire virtual machines can be created with ChatGPT’s help. Virtual machines (VMs) are “computers within computers” in that a physical computer, or the host, houses multiple virtual computers, or the guests - meaning the applications of VM from cloud storage to cybersecurity, are widespread. In this particular case, Jonas Degrave, a research scientist at Google DeepMind and in this case, the software user, connected to various websites through the “mind” of ChatGPT” - something that users can do on a conventional VM as well.

It goes without saying that ChatGPT is becoming a big deal, and not just for the mere graduate student or computer programmer. Rather, it is slowly finding its way into medicine. The software’s ability to rapidly retrieve and extract online data makes it a plausible candidate for medical education, administrative work, and possibly even, medical diagnoses.

When pitted against the United States Medical Licensing Examination, ChatGPT has been reported by Kung et al. to have passed Step 1 and Step 2CK of the exam - both of which are administered to second year and fourth year medical students, respectively. The researchers report that the software “yields moderate accuracy”, “demonstrates high internal concordance” (or consistency), and possesses the “potential to augment human learning in the domain of medical education.”

The Medical Futurist also reports that the algorithm behind ChatGPT does “fairly well in interpreting English-language texts”, making it a suitor for summarizing medical records based on a combination of details, including family history, medical history, and lab results. The article also states that it can be a useful assistant in a clinical context by answering commonly asked questions or booking appointments.

In assisting with charting, ChatGPT seems to be “eerily good” according to Jeremy Faust, MD - an emergency medicine physician at Brigham and Women’s Hospital. Dr. Faust’s request was for the software to write “a ER medical chart about a patient who has a cough. Age: 60. Sex: male. The differential diagnosis is covid vs pneumonia vs an unspecific viral syndrome. The X-ray is negative. The Covid test is pending. He has normal oxygen levels, normal vitals. By the way, he smokes. He is safe for discharge.`` The product was a “completely reasonable” chart that exhibited good guidance and fair thinking.

Nevertheless, this system is not without flaws. For one, it has come under scrutiny for producing articles that do not exist. In the same article, Dr. Faust asked it to make a differential diagnosis based on a limited description of a pretend patient. The most concerning part was when “Dr. OpenAI” was asked to defend, with literature, why oral contraceptives cause costochondritis - an unfounded claim. Dr. OpenAI’s response was a fabricated study, but using real peoples’ names and a real journal’s name.

Thus, it stands to reason that ChatGPT may not be entirely ready for a full-on integration into medicine. There are merits to it that will ultimately help physicians cut down on administrative work, but there are also issues regarding its usage in more rigorous, intricate tasks. Concerns over misinformation and inaccuracies will inevitably exist, but increased investment and continued research into ChatGPT’s functionality will make it a powerful and beneficial tool that can best support medical staff.

Many have already proclaimed that the advent of ChatGPT marks the beginning of a new technological revolution. It will only be a matter of time before the public truly realizes ChatGPT’s direct impact in furthering healthcare.


  1. “6 Potential Medical Use Cases for CHATGPT - the Medical Futurist.” 6 Potential Medical Use Cases For ChatGPT, 19 Dec. 2022,

  2. Bromley, Joe. “I Used CHATGPT to Research My Dissertation — Here’s Why It’s Fine.” I Used ChatGPT to Research My Dissertation — Here’s Why It’s Fine, 31 Jan. 2023,

  3. Degrave, Jonas “Building A Virtual Machine inside ChatGPT” Building A Virtual Machine inside ChatGPT, 3 Dec. 2022,

  4. Doshi, Rushabh H. “Promises - and pitfalls - of ChatGPT-assisted medicine.” Promises - and pitfalls - of ChatGPT-assisted medicine, 1 Feb. 2023,

  5. Shankland, Stephen. “Why We're All Obsessed with the Mind-Blowing Chatgpt AI Chatbot - CNET.” Tech Computing Why We're All Obsessed With the Mind-Blowing ChatGPT AI Chatbot, 17 Feb. 2023,

Post: Blog2_Post
bottom of page