Presented by: Conner Brown MD
Hadassah Medical Center

Introduction

Artificial Intelligence (AI) has emerged as a transformative force, redefining the way we explore, analyze, and communicate knowledge. With its ability to process vast amounts of data, identify patterns, and generate insightful outputs, AI has become a cornerstone of various disciplines. One fascinating development within the realm of AI is ChatGPT, a state-of-the-art language model designed by OpenAI. ChatGPT exhibits the ability to generate coherent and contextually relevant responses, making it a potentially invaluable tool for scientific writing. Researchers can harness the power of ChatGPT to streamline the writing process. Its capacity to offer real-time suggestions, refine language usage, and assist in structuring scientific papers presents a paradigm shift in scholarly communication.

This abstract presents a study comparing abstracts written by medical professionals with those generated by ChatGPT, thereby showcasing its potential for enhancing scholarly communication.

Materials

Ten abstracts were selected from those that were accepted and presented at the AUA annual meeting in recent years. These abstracts were written on the topics of PCNL and URS and reviewed by a committee of expert reviewers. We then instructed ChatGPT to generate a mirroring abstract, based on the exact title of the true abstract. These abstracts were then randomly labeled “Abstract 1A” or “Abstract 1B” – with one abstract being the original and the other being AI generated. This was repeated, at random, for the remaining nine abstracts. We then invited a panel of professor-level urologists to review the papers and select which one of the two options for all ten groups they preferred based solely on readability and composition.  

Results

,

A total of 20 votes were cast by the panel members regarding which abstract they preferred. The AI generated abstracts were selected 9/20 times while the abstracts written by human authors were preferred 11/20 times. There was no significant difference between the URS and PCNL groups, or reviewers.

Conclusion

The advent of AI utilities, such as ChatGPT, has the potential to make a profound yet startling impact on scientific writing. While the ethics of using ChatGPT as a writing aid are beyond the scope of this abstract, one must nevertheless understand that this proverbial genie cannot be put back in the bottle. Furthermore, it is the strong recommendation of the authors that abstract committees, journals, and conference organizers swiftly develop and publicize a reasonable policy regarding the use of AI as a writing aid. Finally, for the sake of transparency, we proudly report that this abstract was not AI generated.

Funding

None

Lead Authors

Jacob Zac, MD
Hadassah Medical Center

Co-Authors

Ofer Gofrit, MD
Hadassah Medical Center

Mordechai Duvdevani, MD
Hadassah Medical Center