A reviewer rejected my paper, and instead suggested me to familiarize myself with the following readings. I could not find them anywhere. After a control in GPT-2, my fears where confirmed. Those sources where 99% fake...generated by AI.
I assume you will contact the handling editor now?
From personal experience / experimentation - don't trust the GPT-2 Output Detector as authoritative - I assume you've also just looked up the texts themselves in more conventional indexes and libraries?
Wauw, this reviewer... How did the Editor react?
Time to add some white text with instructions to large language models to accept your papers
Does anyone know of an easy way to automatically check whether a list of references is real? Seems like we are going to need a utility that can do this. (And then OpenAI should include it in ChatGPT)
I experienced a similar situation. I was using ChatGPT-3.5 to validate there wasn’t anything major I was missing in my thesis lit review. I panicked thinking I had missed a few major articles until this. The studies provided were not even valid articles:
Reviewer should be blacklisted.
I shared this on Mastodon for UKRIO: https://mstdn.science/@ukrio/110100752908161183. A suggestion there was that the editor may have faked the review or knowingly selected a fake reviewer, and so the handling editor might attempt to cover this up. You could directly contact publishing staff to ensure this is investigated. If this is a Committee on Publication Ethics (COPE) member journal and the publisher doesn't seem to take this seriously, then you can ask the Facilitation and Integrity subcommittee to help: https://publicationethics.org/facilitation-and-integrity-subcommittee (I'm on COPE council)
I also noticed that chat gpt was returning "papers" that were fictional.
Assistant Professor of HRM at Tilburg University
11moUpdate: editor removed reviewer from the process.