Fact vs fiction

FigureArtificial intelligence (AI) is a disruptive technology that challenges traditional ways of doing work. The publishing industry is no exception. Like using a calculator to quickly solve math problems, AI is being adopted by authors as a tool in professional writing. AI can improve composition for authors who struggle, yet can also introduce significant risks. At a large meeting of nursing editors from a broad range of journals, views of AI use in scholarly writing ranged from disdain to recognition of its value if applied ethically. Editors from academic settings realized their students use AI to compose some or all of their papers. Many editors received journal submissions that suggested AI use. How to respond to evidence of undisclosed AI use led to more questions than answers. As Editor-in-Chief of Nursing, my primary focus is the accuracy and integrity of published content. Information in the journal is used to inform nursing practice in direct patient care and in other professional nursing domains. AI has introduced new challenges into the manuscript review process. Author guidelines for Nursing, like most nursing journals, specify that any use of AI in manuscript preparation must be disclosed. Authors are accountable; however, compliance is not always easy to determine. AI can introduce strange or awkward word combinations in place of established terminology, termed tortured phrases, that may indicate AI-generated content. An example is “orthopedic vitality” rather than “bone health.” Tortured phrases can be edited, but the bigger concern is hallucinations—AI-generated content that is patently false but may come across to readers as authoritative and accurate. False information is the clear and present danger for obvious reasons, especially in clinical articles. Hallucinations can even impact the reference list and generate references that do not exist in the real world. Expert peer review and thorough fact and reference checking are vital to assure content integrity. These issues are not theoretical “what if” musings about the hazards of AI. Like most other professional and scientific journals, we receive some submissions with tortured phrases and hallucinations, including nonexistent references. Our human Nursing editorial team is diligent in efforts to respond and preserve content integrity. Likewise, nurses must validate any content they derive from AI-generated sources. This validation step is critical. Otherwise, nurses can be left accountable and “holding the bag” if the foundation for their actions that produce an untoward outcome turns out to be a hallucination. LINDA LASKOWSKI-JONES, MS, APRN, ACNS-BC, CEN, NEA-BC, FAWM, FAAN EDITOR-IN-CHIEF, NURSING2025