I guess I'm just stuck on understanding what's problematic about using this technology to do research and writing apart from its tendency to make occasional embarassing errors.
Comments
Log in with your Bluesky account to leave a comment
I think it's too difficult to have this conversation online. but perhaps i will write up my thoughts. but my main position is that there are some kinds of academic work that we want people to do. and when people don't do that work, we lose the primary thing we care about.
I think it’s about not allowing efficiency arguments to trump all other values. A tale as old as time, and one that is sometimes co-opted for reflexively anti-progress views. I understand the suspicion. I think this topic would make for a great symposium.
You both probably already saw this, but if not. Despite the title, I actually thought the survey results showed a lot of agreement on a number of permissible/impermissible uses https://www.nature.com/articles/d41586-025-01463-8
interesting when i started writing my conclusion was the author had been sufficiently destabilized by technology that it was cause for attention; now i think the foundation is evaporating. no normative take
One doesn't have to think it's problematic/relevant to acknowledge that some other people may find it so. I'm inclined to think that under those circumstances, the onus is on those who resist disclosure to say why, as with disclosure of non-bioengineered/organic/free-range/etc
I think a lot of food labeling is about developing common understandings, common lexicons (food identity standards, common and usual names) and at a time when the common understanding of authorship is being disrupted, I think provisionally crystallizing some disclosures makes sense
Putting your name on something as author is a representation after all, and we should try to preserve/evolve common understandings of what that representation means
I wish that’s what food labeling was about! Alas I think a lot of food labeling is about trying to capture market share by spreading misunderstanding & playing into people’s fears about purity & quality.
I think you're saying disclose if it's not used for certain authorial tasks and I think that sounds like a good proposal too--but I don't think I understand the proposal to have work segregated in a journal.
I mean to say that we can't possibly create a disclosure rule for all things that some percentage of the reading population want -- disclosure is a tax! So why not start by seeing if there is an actual appetite for this by having a journal of pure writers and see if it gets traction.
One could! Just not sure why that bars others from disclosing that no-AI was used in something they authored. We also discussed this at our school regarding AI used in writing student recommendations. Some of us raised whether we should consider disclosure to students and judges if AI was used
I don't think it would be problematic if recommendations had a checkbox asking if AI was used. But my proposal is why can't someone writing a recommendation include a note declaring no-AGI was used in the drafting.
Comments