Large language models and scientific publishing

ChatGPT and similar large language models (LLMs) have rapidly caught the scientific and popular imagination, inviting reflections on their potential and risk for science (Grimaldi & Ehrler, 2023; Nature, 2023; Thorp, 2023; van Dis et al., 2023).

As editors of JDIS, we have decided that a software program cannot be an author, because it (or its makers) cannot take responsibility for the integrity of scientific work written with the help of this tool.

Yet, we tend to agree with van Dis et al. (2023) and Grimaldi and Ehrler (2023) that these tools will inevitably be introduced to scientific writing and publishing. In fact, LLMs may soon become integrated in the process of writing articles, and even peer review and editorial decisions.

For this reason, and following suggestions made in (Nature, 2023) and (van Dis, 2023), we will for now require all authors of submissions to JDIS to acknowledge the use of an LLM if they did so.

Like many colleagues, we are following the developments involving LLMs with great interest and excitement about their potential to accelerate and improve the scientific publishing process.

Co Editors-in-Chief

Ronald Rousseau & Liying Yang

Vice Editors

Johan Bollen & Zhesi Shen