That's the first time I've been asked by a journal not to use chatGPT when reviewing an article. Not sure it occurred to me as an option - has anyone given it a go to see how good is the review?
Comments
Log in with your Bluesky account to leave a comment
My not so serious question would be whether a JokeWriter 2000 (TM) pen would be allowed as a tool to write the review. Terry and Andy wrote a whole book with it.
Jeez. That’s worrying! I did an experiment running year 1 essays through it and asking for grades and feedback. After that, I wouldn’t ever recommend it for real grading of UG work, so I really wouldn’t recommend it for journal reviews.
I have to confess, I have seen a couple of approaches over the years that were broadly equivalent to doing this. And I’ve certainly seen feedback that’s as general and vague.
I'd have serious concerns about it given that by default OpenAI collects uploads and inputs. I would take the view that it would be inappropriate of a journal or reviewer to expose the material to reuse, so I'd agree with this journal's policy :) I'm sure this scenario happens often, though!
ChatGPT excels at improving grammar, clarity, and organization in journal papers but lacks domain-specific expertise for deep analysis. It’s best for polishing drafts, ensuring readability, and checking formatting, not for validating methodologies or assessing research impact.
Comments
I've said to a couple of students that they can use AI to write work if we can use AI to mark it...