To chat or not to chat, Generative Pre-trained Transformer?

Abstract

The baton has been passed on. A new editorial board is in place at JIGIMS. I sincerely thank all those members of the institute who have posed faith and confidence in me by entrusting me with the responsibility of the new Editor-in-Chief of the JIGIMS. It is an honor, and I must also acknowledge the past Editors-in-Chief who have nurtured this journal. I am fortunate to have a passionate team of editorial board members who collectively possess a vast experience of working on editorial boards of various indexed journals. The team is working with clear targets in sight although it is a bit challenging in these changing times when many new journals are being launched frequently. The latest player in scientific publication is the use of artificial intelligence (AI)-assisted technologies. AI is now being used both by the authors and the publishers. The authors may utilize it for journal selection, identifying plagiarism, language editing, and proofreading. The journals are benefitting by finding peer reviewers and in various other editorial tasks like checking statistics. The matter gets complex when the issue of the use of AI in hypothesis generation and writing manuscripts is raised. The National Medical Commission (NMC) in India has introduced guidelines which include a publication requirement for the promotion of medical faculty.[1] This has also resulted in unique burden on faculties, medical institutions, and journals. There is a deluge of manuscripts being submitted to the journals which creates a crisis of sorts in finding good reviewers. This, in turn, may affect the quality of publications. A large number of workshops are being conducted on research methodologies by various institutions and organizations. There is no doubt that in the long run these measures will help to improve the overall environment of research and publication in the country. However, the need of the hour is to include discussions on the techniques of peer review and rationale use of AI in scientific publication during these workshops. These are the two areas that have the potential to complicate the research environment of the medical institutions in our country. The faculties of our medical institutions are also the peer reviewers of journals all over the world. Training a new crop of reviewers might preempt the taking over of the review process by AI because, after all, the science of current AI cannot replace the art of an experienced peer reviewer. Large language models like the Generative Pre-trained Transformer (GPT) model do not fulfill the International Committee of Medical Journal Editors (ICMJE) criteria for authorship. However, recently, it has been advocated to acknowledge AI as a contributor to research[2] and ChatGPT be accorded the authorship attribution.[3] AI in scientific publication is here to stay and flourish. But can it replace human creativity and comprehension that is the hallmark of scientific publishing? After all, their prowess is derived from the creativity of humans in the past only. And the moot question, are we going to witness the triumph of science of AI over the art of scientific writing? More thoughts on this are shared by Chandra[4] and Dharwadkar[5] in this issue. These are interesting times in the world of scientific publishing. It is time to amalgamate the art and the science so that they join hands for the development of scientific publication by increasing creativity and decreasing the time consumed. Perhaps the answer lies with the regulatory bodies and the time has come for ICMJE and NMC to formulate guidelines!

Ähnliche Arbeiten