Artificial intelligence (AI) is not only enhancing how we diagnose and treat patients but it is also impacting the publishing landscape. From content creation, publication, distribution to marketing, AI is markedly changing the publishing industry.
Over the recent past, technological evolution, particularly AI has impacted scholarly writing as well as publishing. Use of AI tools has allowed some automation of various aspects of the manuscript writing process such as spelling, grammar, style and plagiarism detection. These AI programmes are time saving and they also improve the overall quality of the scholarly writing process.
However, recently AI chatbots have made their presence felt in the publishing world. The best known of these is ChatGPT from OpenAI (San Francisco, USA). However, more recently AI chatbots from Google (BARD) and Microsoft (Bing Chat) have also been launched. ChatGPT was first launched in November 2022. It is an open source, natural language processing tool, designed using human feedback, is freely accessible and the platform has already attracted millions of interactions. It answers questions and converses in an eloquent and well researched manner. When asked a question it will generate a response quickly (30sec). It became hugely popular in writing school essays but ChatGPT has also been used to formulate scholarly manuscripts and it can even provide references. (1)
This has led to wide ranging conversations regarding its implications in the scientific publication field. There are numerous other concerns about copyright, plagiarism and authorship of text produced by an AI Chatbot. Apparently text generated by AI is not perceptible to anti-plagiarism software. There are several scientific and medical journals who have released statements and policies on how and when these AI Chatbots can be used in the writing of a manuscript (2) Many have defined policies to guide the use of large-scale language models in scientific publication, prohibiting naming of such tools as an “author” due to the lack of these tools taking responsibility. It advises scholars and researchers who use these tools to document the use of AI chatbots in the methods or acknowledgment sections of manuscripts. Elsevier, which publishes a range of peer-reviewed journals, released a new policy governing how these technologies can be used and have stated that AI cannot be listed as an author and its use must be properly acknowledged. (3)
In South Africa, the role of Chat GPT in the teaching environment as well as in scholarly writing and publication, has attracted interest and there have been many webinars and meetings discussing the value and role of these AI chatbots. Due to the functionality of ChatGPT there is some consensus on the growing necessity of implementing AI author guidelines in scholarly publishing. As far as I am aware there are no specific guidelines issued by the Academy of Science of South Africa or any other university or scientific entity. The editorial team at Wits J Clin Med is of the opinion that as an author of a scientific manuscript, one takes responsibility for your manuscript or section of the manuscript (if multi-authored) and one also has to be responsible of the final full manuscript. AI Chatbots do not have this capacity and thus should not qualify as authors. Thus, Wits J Clin Med policy is that AI Chatbot-assisted manuscripts must carry an acknowledgement regarding its use in any section of the manuscript. We acknowledge that AI field is evolving rapidly and may need to refine the current guidance on AI Chatbot assisted manuscripts.
Pravin Manga
Editor
Wits Journal of Clinical Medicine
