Writing and Research Tools
The landscape of academic work benefits from AI applications. It uses a set of solutions designed to streamline research and writing workflows, ranging from article summaries to supporting complex systematic reviews and meta-analysis.
Article review and summary
Popular AI tools like Notebook LM and ChatGPT are available immediately for tasks like article reviews and summarization.
These platforms allow users to upload documents and generate a summary or critical review of their content. Although it may seem convenient at first glance, there are significant caveats to use in scientific reviews, especially with regard to privacy and security.
When performing scientific reviews of articles, the fact that most journals do not currently allow AI tools to be used can raise a major dilemma.9 This limitation is attributed to at least one basic concern. When an article is uploaded to an AI service like OpenAI (which drives ChatGPT), users effectively lose control of that information, leading to potential public disclosure of unpublished sensitive data.
The risk is that author manuscript data, which is intended to maintain confidentiality during the peer review process, can be inadvertently used to train AI models in the future. Such scenarios have already sparked many lawsuits over the misuse of data to train large-scale language models. This is the reality that underscores why surgeons should avoid such breach of confidentiality when using these tools.
Systematic review
Tools available to perform complex research tasks such as systematic reviews are becoming increasingly sophisticated. Specific examples include Swift-Review, Rayyan, and ResearchRabbit, which is one of the best-developed tools.
An example of the possibilities of a systematic review solution is derived from Abigail Loszko, MD, who recently conducted systematic reviews simultaneously, using traditional methods, and the other using only AI tools.10 Although AI methods were not as thorough or well developed as traditional approaches carried out by research librarians, this study demonstrated significant time savings using AI tools. Furthermore, AI solutions identified a reasonable number of additional papers that were not found using traditional search methods, highlighting the possibility that AI tools could augment, rather than simply replacing established research practices.
Meta-analysis
Moving to even more complex research tasks, AI tools are also emerging to perform meta-analyses such as Elicit, Grapha, and Datasquirrel. However, meta-analysis is significantly “difficult” as it requires AI tools to extract, pool and analyze data from multiple sources. As a result, most of these tools are still widely available.
Given the substantial amount of manual data monitoring that still needs to be performed to produce reliable results, it is now safer to perform these analyses manually. Nevertheless, the rapid pace of development suggests that these tools will not be done soon before they are guaranteed to be incorporated into everyday surgical research workflows.
From ambient AI to predictive tools and sophisticated research support, the application of AI in surgery is just an example of the potential for transformation of this technology. Despite certain ethical considerations and current provisions from journals that may require disclosure of AI use, these tools are evolving rapidly despite provisions from journals that may ban them for scientific reviews. AI quickly becomes ubiquitous in our professional lives, just like the spelling and grammar checkers we rely on every day. The most common word processing tools already incorporate predictive text capabilities built into AI frameworks.
AI tools represent the future of surgical practice and research. Undoubtedly, familiarity with them will make certain aspects of your professional life easier. It is of paramount importance that surgeons remain keenly aware of the pitfalls, potential mistakes and important ethical considerations associated with use.
Thoughtful recruitment, coupled with an understanding of those limitations and responsibilities, is key to enhancing patient care and exploiting the full potential of AI in streamlining surgical workflows.
