In this editorial we report on an experiment to use generative artificial intelligence (AI) to write a retrospective editorial on the first decade (1985-1994) of the Australasian Journal of Educational Technology (AJET). Amidst growing uncertainty about the ethical use of AI in academic publishing, in this "meta-editorial" we explore the capabilities and limitations of current AI tools for complex scholarly writing. We tested four mainstream AI tools (Google Gemini, ChatGPT, Microsoft Copilot, and Claude) using identical prompts. Each tool produced outputs of varying quality, structure, narrative coherence, length, and accuracy, with issues ranging from incomplete analysis to fabricated references. No tool was able to produce a manuscript of publishable quality without the need for substantial human editing and fact-checking. The experiment revealed valuable insights into AI’s current capacities, limitations, and implications for academic publishing ethics. The findings underscore the importance of transparency, disclosure, and critical engagement when integrating AI into scholarly writing, including how such tools may influence both research practices and students’ learning experiences. The editorial concludes with a call for clearer, universal guidelines for the ethical use of AI in scholarly publishing to maintain the quality and integrity of academic discourse.
Corrin et al. (Tue,) studied this question.