@SerenaJ@historians.social @L_howes @histodons
Serena, I think your comment is exactly on point. But verification of references has been a problem before, and it is up to us to advocate for an academic culture that will no longer tolerate such misrepresentation (lies, actually) – regardless of whether the source is a poor bot who is trying to do its best, or whether it is a human who is taking deliberate liberties with truth. After all, the author is fully responsible for their words, regardless of how the those words were generated.
I have written on facticity, authorship, and academic integrity in the Sentient Syllabus Project e.g. here https://sentientsyllabus.substack.com/p/generated-misconduct
You are right that the problem is that verification does not scale. But we can make it more scalable. A simple first step would be to require a collection of quoted text to accompany a manuscript submission (could be a photograph of a historic document), with a statement what precisely is being referenced. Plus, a link to the source if it is online. That is easy to do – for an real reference, and for a generated reference, requiring this would force authors to verify that the paper actually exists; a welcome side-effect might be that such verification is likely more work than getting an actual reference in the first place.
As a corollary, I think we should deemphasize paraphrasing and teach our students to quote instead.
Of course there would still be those who would commit fraud, but I would assume those would be in the vast minority.
Thank you!
🙂
