Recording a historical phenomenon in the editorial quality control of an editorial manager: a 'manuscript' file is not explicitly needed for manuscript submission

Dear Publishing Research Editors,

The core of an academic paper is the main body of text, herein referred to as the 'manuscript file'. Many journals employ human-managed but artificial intelligence (AI)-driven tools or platforms, the editorial managers (EMs), onto which academic papers, including the manuscript file, and associated files (e.g., letters, declarations, title page, conflicts of interest statements, data supplements, etc.) are uploaded to complete the submission to a journal^[1]. Usually, and at minimum, basic mandatory files would include the manuscript file and ethics declarations, both of which are essential. Thereafter, depending on the level of initial quality control by that journal, uploaded files are verified by humans for compliance (e.g., ethical, stylistic, etc.). Following file verification and technical checks, editorial handling and peer review then begin.

AI is playing an increasingly greater part in academic publishing, including in ways that challenge human endeavor and basic ethics (e.g., of authorship principles)^[2]. Evidence of errors or failed quality control will thus be required for systems in which there is a human-AI interaction to appreciate where faults exist, and who is responsible for them, i.e., AI or humans. EMs thus serve as an ideal system to scrutinize because their functionality tends to be controlled by AI, whereas the output (in this case, uploaded files or declarations) is then verified by humans. There is an ongoing exercise to appreciate weaknesses and problems with EMs that might allow for their abuse by authors for deceptive practices, such as their role in fake peer review^[3], or the lack of stringent quality control by human journal managers, thereby reducing the level of trust in or efficiency of these platforms. This letter records 15 cases of EMs of 14 ranked and indexed Elsevier journals, all of which employ Aries Systems Corporation's Editorial Manager^®, noting how the requirement for a manuscript file was not explicitly stipulated as a mandatory item, following a chance discovery during submission to those journals between 31 January and 22 April 2023 (Fig. 1). Approximately one year later, on 13 April 2024, when those EMs were verified again, this glitch no longer existed, i.e., a manuscript file is now explicitly indicated as mandatory (evidence not shown in Fig. 1).

Figure 1. 

Fifteen examples of editorial managers used for manuscript submission to 14 Elsevier journals for which submission of the manuscript file was not indicated as a mandatory item, discovered during submissions by the author to these journals. (a) Educational Research Review; (b) The Leadership Quarterly; (c) Journal of Plant Physiology; (d) Biochemical and Biophysical Research Communications; (e) International Journal of Information Management; (f) Diabetes & Metabolic Syndrome: Clinical Research & Reviews; (g) Plant Stress; (h) Journal of Business Venturing; (i) Progress in Neurobiology; (j) Emotion, Space and Society; (k) Neuroscience & Biobehavioral Reviews; (l) Journal of Business Research; (m) Journal of Plant Physiology; (n) Early Human Development. Editorial manager: Editorial Manager® (Aries Systems Corporation). Data collected (i.e., discovery and accompanying screenshots) between 31 January and 22 April 2023. Screenshots (Fair Use) trimmed to include only basic relevant information, so not all items in all drop-down menus are displayed. In one case (e), the author had already uploaded the mandatory files when the screenshot was taken, explaining the green text and ticks not present in the other screenshots (in which mandatory items are listed to the left of the drop-down menu as red text and bullets). In two cases (j, k), 'Required for Submission' is missing. Even though (c) and (m) are the same journal (Journal of Plant Physiology), the content of their drop-down menus changed considerably between sampling dates: 27 February 2023 and 16 April 2023, respectively.

This observation reveals two aspects of automated online submission platforms. The first pertains to the lack of EM management by humans. All journals in Fig. 1 are likely to have received high submission volumes when a manuscript file was not explicitly stipulated as a mandatory item, suggesting that authors would have ignored this glitch. The existence of this glitch also suggests that the journals' editors were not aware of it. The second issue is procedural, i.e., despite a manuscript file not having been indicated as a mandatory item, authors would still have submitted one. In other words., the 'voluntary' status of the submission of a manuscript file would likely have been ignored by submitting authors, simply because the submission of a paper without a manuscript file would not make any practical sense. Even if authors complied with the instructions implicit in the non-mandatory status of the manuscript file, i.e., even if the authors did not upload a manuscript file, the submission would likely still have been possible. However, in this case, human verification post-submission would have detected the absence of a manuscript file, and authors would have likely been alerted of its absence, and requested to upload a manuscript file.

What then is the value of this exercise? As was noted above, there is a historical need to formally record evidence of EM mismanagement, so that in the future, it can be appreciated whether humans or AI, or both, failed. In some cases, journal submission mismanagement can be extreme^[4]. It is not clear precisely when the item menus of the EMs of those 14 journals were corrected between 2023 and 2024. However, if any author were to observe them today (in 2024), they would never know that this issue existed, hence the importance of the evidence presented in Fig. 1. Whereas formal errata and corrigenda exist for authors to register errors in their published papers, there is no formal mechanism to keep a historical record of publishers' errors, like the EM glitches recorded in this letter, so papers such as this one serve a dual historical and accountability purpose.

Authors who employ EMs for manuscript submission need to be constantly vigilant of glitches and errors in these systems, because they are fundamental tools of the publishing industry, and because their submission experience will be negatively impacted as a result of such issues. More importantly, when discovered, authors should make a concerted effort to record such glitches and errors, for example as case studies in papers like this, to serve a historical purpose, and to hold publishers accountable. Editors also need to pay closer attention to the EM that they employ for their journal/s. Finally, publishers need to responsibly correct their EM when errors and glitches exist, as Elsevier did in this case.