Indications of LLM use in writing peer-reviewed psychology manuscripts

Studies have found evidence of LLM use in the writing of scientific publications after the launch of ChatGPT in November 2022, with studies finding evidence of excess stylistic word use favoured by LLM models (see Kobak et al., 2025; Liang et al., 2024). These studies have compared publication abstracts pre- and post-ChatGPT launch in order to identify the stylistic word choices made more frequently by LLMs than human writers. These excess style words include words such as 'delve', 'intricate', 'pivotal', and 'underscore' and can be seen as the fingerprints of LLM use in research publications. In this study, we aim to utilise a list of 28 excess style words identified in previous studies as being linked to LLM use and examine whether a significant increase in the use of these words can also be found in the research field of psychology. In addition, as the use of LLMs have been found to be linked to certain publication outlets (e.g. MDPI, Frontiers), to non-English native speakers, and to lower impact journals (Kobak et al., 2025), publication descriptives such as affiliated country, publication outlet, and journal statistics will also be examined in association with AI-word frequency. References: Kobak, D., González-Márquez, R., Horvát, E. Á., & Lause, J. (2025). Delving into LLM-assisted writing in biomedical publications through excess vocabulary. Science Advanc-es, 11(27), eadt3813. Liang, W., Zhang, Y., Wu, Z., Lepp, H., Ji, W., Zhao, X., ... & Zou, J. Y. (2024). Mapping the increasing use of LLMs in scientific papers. arXiv preprint arXiv:2404.01268.