Regardless of whether the papers are any good (I’d guess not), the flood of AI papers also presents the journals’ essential review process with essentially insurmountable challenges: Reviewers do this work for free (or for the prestige) and can’t spend an unlimited amount of time reading and correcting sloppily summarized AI papers… only for the author to then feed the comments into the next AI. In the worst-case scenario, this whole situation will lead to AI systems writing papers, which are then peer-reviewed by other AI systems and ultimately published without a single human having even read the entire text. That would spell the end of the purposeful, self-critical scientific process we’ve relied on until now, one that has provided us with at least somewhat reliable findings.

deleted by creator