EXPLORING THE ROLE OF chatGPT IN MEDICINE: VIRTUAL ASSISTANT OR POTENTIAL SOURCE OF MISINFORMATION?| APRIL 28TH, 2023

Launched in November of 2022, Chat Generative Pre-Trained Transformer (ChatGPT)  is an unprecedented generative language model tool, acquiring over 100 million users within 2 months, making it the fastest-growing consumer application in history. While ChatGPT is a product of OpenAI, major corporations like Microsoft and Google have now launched their own versions into the public arena.

Given this extraordinary evolution of technology, health professionals and medical educators understandably have notable concerns and questions. For example, can ChatGPT help address the physician shortage in the future? Will ChatGPT enhance clinical decision-making and improve patient outcomes?  Will Chat GPT be more or less likely to spread health misinformation directly to patients? Will ChatGPT enhance medical education, or contribute detrimentally? Can ChatGPT be used to summarize the medical literature in a more understandable and actionable way?

While these are all important questions that many are asking, there are few clear answers to this complex dilemma. Previously, conversations around artificial intelligence (AI) centered on bioinformatics and big data, machine learning, and quantum computing.  Now, ChatGPT, which provides the opportunity for humans to engage in conversations with a machine, has taken the age of technological innovation by storm and is poised to revolutionize the human experience. 

Specifically, let us dive deeper and look at how ChatGPT can affect the field of medicine. Researchers anticipate that “ChatGPT has the potential to digitize the whole healthcare system” by providing medical professionals with plentiful amounts of information and better patient outcomes. According to the New York Medical College (n.d.), “ChatGPT can act as one’s virtual assistant.” Having chatGPT serve as a virtual assistant includes the potential to provide medical information deemed relevant to someone concerned about their health. With massive data sets implemented within ChatGPT, the question remains-what will be the quality of that advice? As ChatGPT quickly pulls in information from various sources, what checks and balances will need to be put in place to protect the public?  

While ChatGPT may offer various advantages, it will likely need regulation to mitigate disadvantages and risks. The New York Medical College has stated that “the need to balance patient privacy and data protection” is crucial for ensuring that ChatGPT does its job acceptably. 

With regard to the medical research literature, the use of ChatGPT will likely introduce serious flaws, at least for the foreseeable future. In a recent interview with JMIR Editor Dr. Gunther Eysenbach, ChatGPT fabricated a publication, presenting it as a real citation, and further went on to critically analyze it as if it were a study, rather than a review article. 

“At this point, there are more concerning questions being raised than answers. As a medical profession, we will need to proceed thoughtfully as we have in response to other emerging technologies, considering the risks, benefits, and alternatives of each utilization on a case-by-case basis”, notes, Dr. Jennifer Hartmark-Hill, Family Medicine Physician and Associate Professor in the Department of Bioethics and Medical Humanism at the University of Arizona College of Medicine-Phoenix. 

While ongoing challenges will need to be addressed, ChatGPT has revolutionized the world as we see it today. A number of major companies, such as Microsoft and Google, have released their own versions of ChatGPT. 

While ChatGPT may hold the potential to revolutionize the field of medicine, our comprehension of and ability to mitigate unintended consequences is still in a nascent phase, and we should proceed with an abundance of caution. 


References:

Gilson, Aidan, et al. "How Well Does ChatGPT Do When Taking the Medical Licensing Exams? The Implications of Large Language Models for Medical Education and Knowledge Assessment." medRxiv (2022): 2022-12.

Harvard Medicine News. "The History of Artificial Intelligence." Science in the News, 7 Nov. 2017, https://sitn.hms.harvard.edu/flash/2017/history-artificial-intelligence/.

New York Medical College. "Envisioning the Healthcare Landscape with ChatGPT." NYMC News & Events, n.d., https://www.nymc.edu/news-and-events/news-archives/envisioning-the-healthcare-landscape-with-chatgpt.php.

Pearl, Robert. "5 Ways ChatGPT Will Change Healthcare Forever, For Better." Forbes, 13 Feb. 2023, https://www.forbes.com/sites/robertpearl/2023/02/13/5-ways-chatgpt-will-change-healthcare-forever-for-better/?sh=3d51de6e7bfc.


By: Himanshi Kapoor and Dr. Jennifer Hartmark-Hill