This is the authors' abstract. We don't add key points for this paper.
In this paper the author argues that generative artificial intelligence (GenAI) has not created a new form of academic misconduct. Rather, it has taken the least-policed form of misconduct we had, that of cutting corners, and stripped it of the three properties that made it survivable. This ‘sloppiness’ was coupled to effort, bounded to one paper, and attributable to someone who lacked diligence. Fluent prose was once an ‘honest signal’, because it was costly to fake, and the apparatus of peer review built upon that cost. GenAI has broken that signal. Other things have broken too, including the evidence chain. ‘Citation laundering’ is where a fabricated claim, cited once by an author linking it to a real source, acquires a provenance despite it being unreachable by any reference check. The author highlights research in health professions education in which various authors have admitted to citing work they have not read, and argues that current GenAI use declarations are largely uncalibrated instruments. The closest look health professions education scholars seem to have taken at their own reference lists appears to predate these tools, and because syntheses inform guidelines and guidelines inform curricula, a fabricated claim can travel from a ‘prompt’ to what a clinician teaches at the bedside. Because journals’ workflows have not traditionally involved verification, blame for abuses all too often falls on reviewers, who are ill equipped to carry it. Ultimately, the use of GenAI means that much of our cognitive effort needs to be relocated from the act of writing into a deliberate act of adjudication. The author describes how this paper was itself produced using GenAI to bring back effortful, bounded and attributable practice leading to a worked example of disclosure that describes practice rather than certifying intention.
Answers are grounded in this paper and cite their sources.
No takes yet. Share an insight, caveat, or question.
Lynn Valerie Monrouxe (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: