Issue 4 - March 2026 - Editorial
The elephant in the submission guidelines: AI and poetry publishing.
Most of us are using it, in one way or another. In journal submission guidelines it's usually referred to something like this: 'We do not accept submissions generated by, or created with the assistance of, AI.'
I'm reminded of my early days writing haiku when that new-fangled thing called the internet threatened, according to some, to destroy the grounded realism of haiku composed in nature, written in notebooks, printed on paper. Would plagiarism run rampant? Would everyone end up writing sci-fi ku?
The submission guidelines for Quail Eggs do not say anything about AI. I'm not saying they never will, I have not made my mind up yet. I am exploring it, testing it out, being surprised, occasionally alarmed, but mostly intrigued.
I began by chatting with ChatGPT about my personal preferences in tanka. Then I asked it to write a tanka in a style that I would be likely to accept. Finally I asked it to critique its own attempt. It gave a very long reply which I will summarise as: 'This submission would probably not be accepted as it lacks grounding in lived experience.' Which may be AI reflecting back to me my beliefs, or an objective judgment of its own abilities. I'm not sure.
AI generated original poetry is one thing, I suspect it's not there - at least not yet. More open to question is the issue of using AI as a workshopping tool, helping a poet revise what they have written. I asked it to suggest revisions to one of my own rejected tanka, taking into account the preferences I had informed it of earlier. It did much better at this. Especially after I asked it to be less agreeable in its responses and more of a critical friend. As in any workshop I was accepting some suggestions and rejecting others, it was responding with alternative suggestions. As in any workshop, I was still the author, it was still my poem. Is this 'the assistance of AI'?
I won't be sending that revised poem out to any journals that have a 'No AI' policy as a courtesy to them, but that depends entirely on trust. I do wonder how many poems that have been in an AI chat in some way are being published, but there is really no way to know.
I suppose what I am pondering is, if a poem resonates with an editor enough for them to select it, however it came to be, then why shouldn't it be published? Do those of us who are editors trust our own judgment enough to deal with that?
Another, more whimsical question occurred to me: will AI ever take the initiative and send a submission of its own to a journal?
My main conclusion on AI so far - and this is really not surprising at all - is that much depends on how you use it, what you tell it, what you ask it, how you train and control it. I think AI could be a useful tool in some ways, but how useful it is depends largely on the person using it in an appropriate way.
In the end, what gave me the deepest doubts about the potential of AI for poetry was when I asked it if there were things that AI definitely cannot do and one of the things it listed was - care.
I would love to hear your thoughts on this topic!
Note: This editorial was written by the editor after a conversation with ChatGPT. Yes, it did offer to write it for me, but I told it that was my job. It agreed.
Comments
Post a Comment