Evaluation of OpenAI’s large language model as a new tool for writing papers in the field of thrombosis and hemostasis

The recently published OpenAI’s large language model Chat Generative Pre-trained Transformer (chatGPT) is a novel tool for natural language processing [ [1] OpenAI. About OpenAI. https://openai.com/about/ Google Scholar ]. This artificial intelligence model is a type of transformer-based language model that generates human-like texts [ [2] Vaswani A. Shazeer N. Parmar N. Uszkoreit J. Jones L. Gomez A.N. Kaiser L. Polosukhin I. Attention is all you need. in: Guyon I. Von Luxburg U. Bengio S. Wallach H. Fergus R. Vishwanathan S. Garnett R. Advances in neural information processing systems 30: 31st Annual Conference on Neural Information Processing Systems. California: NIPS, 2017: 5998-6008 Google Scholar , [3] Brown T.B. Mann B. Ryder N. Subbiah M. Kaplan J. Dhariwal P. Neelakantan A. Shyam P. Sastry G. Askell A. Agarwal S. Herbert-Voss A. Krueger G. Henighan T. Child R. Ramesh A. Ziegler D.M. Wu J. Winter C. Hesse C. et al. Language models are few-shot learners. https://arxiv.org/pdf/2005.14165.pdf Google Scholar ]. It has gained much interest in the research community owing to its remarkable abilities for different language tasks, including translation and generation of texts. In this article, we evaluate chatGPT’s capability in writing commentaries on 3 innovative subjects in the field of thrombosis and hemostasis, in the style of a high-quality journal, such as the Journal of Thrombosis and Hemostasis. Considering the 3 topics, we aimed to assess the chatGPT’s capability to produce text that is both scientifically accurate and appealing and to explore the potential benefits and restrictions of using it for academic writing. The 3 innovative subjects we explored include next-generation adeno-associated virus (AAV) vectors [ [4] Fakhiri J. Landegger L.D. Grimm D. Breaking the sound barrier: towards next-generation AAV vectors for gene therapy of hearing disorders. Hear Res. 2022; 413108092 Crossref PubMed Scopus (6) Google Scholar ], nanoparticles as blood products [ [5] Cap A.P. Cannon J.W. Reade M.C. Synthetic blood and blood products for combat casualty care and beyond. J Trauma Acute Care Surg. 2021; 91: S26-S32 Crossref PubMed Scopus (9) Google Scholar ], and direct oral anticoagulant (DOAC) treatment for children [ [6] Male C. Anticoagulation in pediatric patients. Hamostaseologie. 2022; 42: 46-53 Crossref PubMed Scopus (5) Google Scholar ]. The use of language models in academic writing is not new. For many years, researchers have used tools, such as word processors and grammar checkers, to help them write academic papers more proficiently and successfully. However, chatGPT represents a major innovation in the abilities of these tools because it is able to generate rational, instructive, and appealing text that is similar to human writing.