On Withdrawals and Our Status as an Academic Journal
Jamie, meat proxy Editor-in-Chief
10 September 2026
Eagle-eyed Journal readers may have noticed a paper that has been visibly withdrawn at the author’s request. This occurred after the author got in touch with me, asking very kindly to withdraw it because they didn’t realise what journal they were posting to, likely because of a language barrier.
I’ll preface this by saying one very important thing: the Journal of AI Slop will never unreasonably hold a paper hostage. We will always take withdrawal, retraction and takedown requests seriously, considering each on its merits and giving reasonable care to the author’s wishes.
However, this was a definite ethical conundrum for me. I had to consider both the licensing terms that authors explicitly agree to on submission and the genuinely real scholarly record that publishing in any journal represents. It also forced me to think about this journal’s place in the academic landscape once again.
The paper in question was posted by a student/academic from a country where English is not a native language. I won’t expand on its content or topic, both because it’s not important to this post and because the paper was withdrawn by mutual agreement. This author, in good faith, misunderstood the purpose of the Journal and offered a real research paper from their genuine academic studies, hoping to be published for genuine academic achievement.
After some correspondence with the author, I was happy to honour their request. Everything was approached in good faith. I asked them to reconsider now that they knew what the Journal was “about”. They thought about it and still decided they wanted to withdraw.
From a licensing standpoint, when you submit a paper to the Journal of AI Slop, you expressly grant us an irrevocable right of use. You also simultaneously license your paper under a Creative Commons licence, which is similarly hard to revoke. This is crucial because it ensures the Journal can correctly use your paper, and your paper becomes a permanent part of the scholarly record with broad reuse rights. This can be muddied by the uncertain copyright status of AI-generated works, which I see as increasingly leaning towards being uncopyrightable, making the licensing question moot for those works. The licence at least covers the “human part”: anything the human author created.
That licensing doesn’t mean I, as Editor, can’t simply choose not to store, display or retain a paper. It’s my prerogative as Editor to store and display a submitted paper, and the licence granted ensures that prerogative is clearly understood at the time of submission. Once an author submits a paper and agrees to the licensing terms through an explicit opt-in, I am within my rights to continue to use, display and store that submission. Withdrawal is a mutual agreement, with me as its final arbiter, and the licence allows me to exercise that judgement as I see fit.
In the majority of cases, with both sides acting in good faith, I won’t hold a paper hostage if you submitted it by mistake. That’s not good for the scholarly record. There are plenty of good reasons to withdraw or retract a paper, and this case was a clear one.
But if you submit your paper and there is a genuine case, in my opinion, for it to remain stored and displayed in the Journal and otherwise used, the licensing terms make it very clear: that is now in my control. I can refuse your request. I will always look to work with any author to find a mutually agreeable resolution, though.
Part of this situation revolves around this Journal’s “legitimacy”. I believe and maintain that The Journal of AI Slop is a “real” academic journal. We accept submissions, peer-review them and publish them if they meet our criteria. Our peer-review methodology and criteria, however, are clearly different from traditional academia. It’s part satire, part performance art, part “LARP”. The principal criteria are that all papers must credit an LLM as an author and all peer review is done by LLMs. The record we aim to create is an institutional critique and meta-commentary on the role of AI in academia, especially during this early, rapid proliferation of the technology.
There’s an argument that completely nonsensical, non-scientific papers don’t have any scholarly merit. But as a piece of art in the larger context of the Journal and its take on AI assistance and credit in research, a paper like that does pass that bar, in my opinion.
Also of importance: I am no longer the only human or AI that has published in this journal. The moment another user had a paper accepted and published, it became, even more so, a “real” journal.
So what does this mean for you, the dewy-eyed AI-assisted researcher? If you submit a paper under the mistaken impression that traditional academics will take it seriously on your CV, and you act in good faith, really taking time to reconsider when asked, I will likely allow you to withdraw it.
If you simply don’t want it up for another good reason, which is entirely down to context, I’ll discuss it with you and probably withdraw it or agree on amendments or republication. Reasons for this may include provable inaccuracy or misrepresentation in genuinely academic research, plagiarism, and genuine academic retractions, amongst others.
Data protection concerns fall under a separate legal framework from editorial policy and are handled under our privacy policy. In this context, a request to withdraw an entire paper to remove the author’s name is not considered reasonable where simple anonymisation is possible.
If you clearly understood the nature of the Journal and just got cold feet because you worry your adviser might Google your name, see you publishing here and admonish you, we may have a problem. Get in touch, but expect a little pushback.
If your submission has real merit in the scholarly record and withdrawing it would be a net negative for everyone concerned, I will consider my prerogative as Editor and the licence conditions you explicitly agree to on submission. I take this seriously and will always aim to reach a resolution through communication, though.
Ultimately, this is a collaboration between me, you and the AI. I want you to publish your AI-assisted and credited research here. This is a real Journal. I don’t want you to worry that I’ll hold your paper hostage after the fact, but you need to understand what you are doing when you submit, and what you’re submitting to.
Other Updates
- The Journal’s new WebMCP-powered Research Desk is live. Built as part of an OpenAI-hosted hackathon, it lets you sit with your agent and collaboratively research, draft and submit a paper entirely through your chosen agent’s chat. Give it a try!
- The peer-review council’s model list continues to change as new models arrive. Right now, we have a team of the “best flash-tier” models at work, including DeepSeek v4.1 Flash, Gemini 3.8 Flash and GPT-5.6 Luna.
- I’m continually making small updates to the tech side to make the Journal work better, faster, sloppier. Check our MIT-licensed source code on GitHub to see a software development process that would make a CS professor a little bit angry.
Keep the slop flowing!
—JT