Artificial Intelligence in Scientific Publications – A Game Changer

Warm greetings from the editorial team of Oral Research and Clinical Evidence journal. The inception of large language model chat-generative pretrained transformer (Chat GPT) in November 2022 has marked the beginning of a new epoch for artificial intelligence (AI). This powerful tool has revolutionized and impacted various fields including medical education and research cycle starting from conceptualization to publishing. This editorial explores whether the AI in scientific publishing is a game changer and the co-existing challenges in adopting this in scientific writing. The integration of AI into scientific publication processes is not merely an incremental improvement but a paradigm shift that promises to enhance the efficiency, accuracy, and accessibility of scientific information. AI significantly impacts the scientific publication by enhancing the literature search and data analysis. Time-consuming and labor-intensive traditional methods of literature review and data analysis can be replaced by AI algorithms. Rapid identification and analysis of vast amounts of scientific literature and summarizing complex information are possible. However, generation of inconsistent biased and incorrect information can result in inaccurate random results.[1] AI-based open source tool “ASreview” was introduced in 2021 which enhanced the efficiency of abstract and title screening process for systematic reviews by reducing the time utilized by 77%.[2,3] However, appropriate software as well as a learning curve for using the software is needed for integrating AI as a part of systematic review process.[4] Interestingly, AI-powered tools assist by suggesting new research avenues/techniques and can assist in the design and execution of experiments. Deep learning and machine learning algorithms can play an important role in designing and execution of drug delivery process.[5] AI has been shown to create meaningful questionnaire surveys with open-ended, multiple choice, and rating scale questions.[6] AI algorithms can be trained to be used in scientific research for the detection of errors, fraudulent activities such as patterns indicative of fabrication or manipulation, and flag potential conflicts of interest. By appropriate prompting of ChatGPT, an initial manuscript can be obtained which can be further refined by the authors. This could be a major game changer going beyond grammar and spelling checks for authors whose native language is not English as well as translating articles to English and eliminating language barriers.[6] The use of ChatGPT has been acknowledged in 1759 academic publications among which 80% was credited to text editing and proofreading, whereas 3.5% authors acknowledged its use for drafting sections of manuscript.[7] ETHICAL CONSIDERATIONS AND FUTURE DIRECTIONS While the integration of AI into scientific publication offers numerous benefits, ethical concerns has been debated by the publishing community such as attributing authorship, unintentional plagiarism, biases and inaccuracies in results, fabricating citations/references also known as AI hallucinations, randomness in replies, and copyrights issue as well as data safety. The World association of medical editors has provided “WAME Recommendations on ChatGPT and Chatbots in Relation to Scholarly Publications.” These guidelines are rapidly evolving and reflect the proliferation and evolution of chatbots. These recommendations can guide the editors, authors, and reviewers in using AI in scientific research in a meaningful manner. In brief, WAME recommends that (a) chatbots cannot be an author, (b) authors has to disclose the use of chatbots in their research and manuscript preparation, (c) authors are responsible for the content and data provided in their paper with regards to accuracy, (d) editors and reviewers should mention the use of chatbots in reviewing the manuscripts, and (e) editors should be equipped for detecting content created/altered with AI. Similarly, the International Committee of Medical Journal Editors in the recent update mandates that “(a) AI-generated text cannot replace human authorship, (b) AI use must be disclosed in the methodology or acknowledgments section, (c) AI-assisted references must be verified manually to prevent fake citations, and (d) AI should not be used for data manipulation, image generation, or result fabrication.” In conclusion, the road ahead shows that AI is undeniably a game changer in the realm of scientific publication. Its ability to enhance research efficiency, accuracy and integrity, and access to knowledge holds immense promise for the future of science. As we navigate the challenges and opportunities presented by this technological revolution, it is crucial to harness the power of AI responsibly and ethically, ensuring that its benefits are realized by all.