A student asked how to cite/use AI ethically. Specifically, they asked (and I am paraphrasing here)”
“If I use ChatGPT or other AI tools for brainstorming and making test files – like having it write code for a project to generate some random sequences and introduce an error rate so I have some options to test aligners.”
I’m not sure we have a great answer. What I found so far:
*This paper from Software Sustainability Insitute has links to the items below and expands some sections, I thought that parts 6.4 “Cite any software that you view as having contributed…” and 6.5 “Provide additional details..” were most useful.
*From MLA, the domain of the humanities, is surprisingly helpful – they explain about citing both the source code and as of March 2023, generative AI.
*APA advice is still forming. For now, the answer I prefer is that it depends on if you can save the chat (or share or retrieve it) then treat it as software. If not, treat is as a personal communication.
*Folks have asked ChatGPT to weigh in. Without the date specified, I think those aren’t precise enough.
*And, this guide from MIT is clear about code. Note that this guide is linked to by many others: http://integrity.mit.edu/handbook/writing-code
*Colleagues found this article for me, which I was super excited to dig into. However, after a quick review I think it’s a big fat AI joke.
*Lastly, Nature and Science weighed in and have explicitly stated that AI can’t be an author. I’ll keep looking and update when the answers get clearer.
We’re all working on these questions. Here’s a post published about the same time as I assembled this one, that’s very useful: https://distlib.blogs.com/distlib/2023/01/how-to-cite-chatgpt.html