Use of AI in publishing with JCRS

Abstract

Artificial intelligence (AI) is a controversial topic. Some are all for it, others are wary and skeptical. Whichever side you may be on, AI is being used every day in visible and invisible ways. It is becoming increasingly difficult, perhaps impossible, to avoid it in simple Google searches or customer service queries. The application of AI in scientific publishing presents unique and serious challenges to the integrity of academic scholarship. The conversation moves beyond the philosophical and social debate and crosses into the realm of plagiarism and breach of confidential information. What does appropriate use of AI look like for academic publishing? The use of AI in scientific publishing must support the highest standards of transparency, accountability, integrity, and verification. The Journal of Cataract & Refractive Surgery (JCRS), along with any reputable medical journal, adheres to the guidelines of appropriate use of AI as defined by the International Committee of Medical Journal Editors (ICMJE). Their guidelines clearly state that the use of AI must be disclosed in both the cover letter and the submitted work.1 This includes the use of AI for editorial assistance with formatting, grammar, and stylistic improvements. Using AI to polish and improve a manuscript can save a nonnative English author time and money, but it should only be used to improve syntax and clarity. For JCRS, the use of AI to improve the writing of a manuscript should be disclosed in the last paragraph of the Introduction as well as in a standalone statement at the end of the main text, before the Reference list.2 Using AI as a tool to improve writing is acceptable and appropriate, but AI can never be used as a coauthor or substitute for original thoughts. Following the ICMJE guidelines, JCRS does not allow a chatbot to be listed as an author. A computer program cannot be held responsible for the accuracy and originality of a manuscript.1,3 At the heart of scientific publishing is authenticity and accountability. When an author publishes a manuscript under his name, he is taking full responsibility for the findings of that study. This includes the ability to test, prove, and support his findings, skills of which a chatbot are not capable. Using AI to interpret data is tempting, but can be dangerous. The generated text may be inaccurate, and the wording could be formatted incorrectly or with a particular bias. Of greater risk and concern are AI “hallucinations,” which JCRS editors have encountered in submitted manuscripts. Hallucinations refer to information generated by AI that is nonsensical or impossible but seems to be true. In academic writing, this can manifest as manuscripts that seem well-written with an impressive dataset but in fact are comprised of completely fabricated data and false references. If AI is used to analyze a dataset, the findings must be carefully and thoroughly reviewed. An error published under your name is your error and can have a long-lasting impact on your ability to publish in the future. In addition, patient information and text from a manuscript under peer review or accepted but not yet published is considered confidential information. It is considered a breach of confidentiality to share this information in a public AI search engine. If AI is used in data collection in any way, a disclosure statement should be present in the Materials and Methods section. JCRS does not allow the use of AI in peer reviews. To safeguard the confidentiality of unpublished work and to maintain a high standard of credibility, JCRS's policy for peer review clearly states, “AI tools cannot be used to generate reviews. The journal strongly believes in the integrity of the independent peer-review process, based upon individual subject matter expertise. In addition, generating reviews via AI would require the input of confidential text and data into AI programs as prompts, which is strictly prohibited.”2 Using AI to complete a peer review is a breach of confidentiality and could result in you being excused as a peer reviewer for the journal. AI has transformed the way we live and how we practice ophthalmology. Its application in our preoperative measurement tools and surgical devices is phenomenal. But AI should be avoided in the realm of scientific writing because it does not surpass the human rationale and discretion needed for credible and reputable academic publishing. - JCRS Artificial Intelligence Policy Authors should not list artificial intelligence (AI) and AI-assisted technologies as an author or coauthor, nor cite AI as an author. Authorship implies responsibilities and tasks that can only be attributed to and performed by humans. Authors who use AI tools in the writing of a manuscript, production of images or graphical elements of the paper, or in the collection and analysis of data, must be transparent in disclosing how the AI tool was used, and which tool was used. Authors are fully responsible for the content of their manuscript, even those parts produced by an AI tool, and are thus liable for any breach of publication ethics. Authors must disclose the use of generative AI and AI-assisted technologies in the writing process by adding a statement to the manuscript file in the last paragraph of the Introduction section and also as a standalone statement at the end of the main text. The statement should be placed before the References list, under the subhead “Artificial Intelligence Use” and formatted as During the preparation of this work the author(s) used [Name Tool/ Service] to [Reason]. After using this tool/service, the author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the publication. This declaration does not apply to the use of basic tools for checking grammar, spelling, references, and so on. If there is nothing to disclose, there is no need to add a statement. In addition, authors should not upload an accepted or published manuscript or any part of it into a generative AI tool as this may violate the copyright agreement or licensing terms in effect at the time of acceptance. Artificial Intelligence Tools and Peer Review AI tools cannot be used to generate reviews. The journal strongly believes in the integrity of the independent peer-review process, based on individual subject matter expertise. In addition, generating reviews via AI would require the input of confidential text and data into AI programs as prompts, which is strictly prohibited.

Ähnliche Arbeiten