From Artificial Intelligence to Authentic Writing: Exploring the Prospects and Challenges of AI in Transforming the Landscape of Academic Writing in Hospital Medicine-Slides

Slides

Pdf Summary

This presentation, delivered by Claire Davis, MD, and Matthew Sakumoto, MD, at the 2024 Society of Hospital Medicine Conference, explores how artificial intelligence (AI), specifically large language models (LLMs) like GPT, is transforming academic writing in hospital medicine. The session covers the fundamentals of GPT technology—Generative Pretrained Transformers—and its applications in generating original content, assisting with brainstorming, summarizing literature, and improving writing efficiency.<br /><br />A central case study illustrates a junior faculty member leveraging AI (ChatGPT and Bing) to explore the impact of emoji use in primary-to-consultant physician text paging communication. Using AI tools, he generates research ideas, summarizes prior work, and develops a manuscript, demonstrating how AI can aid various stages of academic writing, from ideation to journal submission. The authors emphasize that AI tools can enhance manuscript preparation by suggesting abstracts, improving readability, formatting references, and even simulating peer reviewer feedback, though human oversight remains essential.<br /><br />Ethical considerations are extensively discussed. The presenters stress transparency in disclosing AI use within manuscripts, underline authors’ full responsibility for content accuracy, and highlight that AI cannot be credited as an author. They review publishing guidelines from journals like JAMA, JHM, NEJM AI, and JMIR, which permit AI assistance with proper acknowledgment but caution against AI-generated content without disclosure due to risks including hallucinations (fabricated information), bias, plagiarism, and confidentiality breaches.<br /><br />AI’s limitations are noted, including its inability to generate genuinely novel ideas, potential for misinformation, and lack of specificity in reviews. Recommendations include rigorous fact-checking, ethical use aligned with journal policies, and the role of AI as a supportive tool rather than a replacement for critical thinking.<br /><br />Overall, the presentation advocates for responsible integration of AI in academic hospital medicine writing to improve efficiency and democratize knowledge dissemination while maintaining authenticity and integrity.

Keywords

Artificial Intelligence

Large Language Models

GPT

Academic Writing

Hospital Medicine

AI Ethics

Manuscript Preparation

ChatGPT

Medical Communication

AI Limitations