Joshua

I recommended rejection for a paper that was, in my opinion, AI slop in both content and style. I was satisfied the editorial decision was for rejection, but it concerns me is that none of the other reviewers seemed to find problematic the AI-typical rhetoric. I don’t want to punish AI use automatically, and, indeed, I use AI in some of my own writing. The ideas content should be evaluated for what they’re worth. But in persuasive or semi-persuasive writing like position statements/perspectives (which this was), narrative reviews, or discussions of research articles, I do think reviewers/readers ought to be on the lookout for typical, whole-cloth AI writing that is super slick. I’m talking about out-of-the-box default model output that is unedited by human and scores 100% on Pangram. It can dress up vacuous claims and poor logic trains with vague clickbait statements and sales-pitch verbiage that inflates the importance of the supporting evidence. You see it all over LinkedIn and it’s begun to infiltrate scholarly journals. Editorial norms are still forming on this issue. What do you think?