Artificial intelligence (AI)-driven language models have the potential to serve as an educational tool, facilitate clinical decision-making, and support research and academic writing. The benefits of their use are yet to be evaluated and concerns have been raised regarding the accuracy, transparency, and ethical implications of using this AI technology in academic publishing. At the moment, Chat Generative Pre-trained Transformer (ChatGPT) is one of the most powerful and widely debated AI language models. Here, we discuss its feasibility to answer scientific questions, identify relevant literature, and assist writing in the field of human reproduction. With consideration of the scarcity of data on this topic, we assessed the feasibility of ChatGPT in academic writing, using data from six meta-analyses published in a leading journal of human reproduction. The text generated by ChatGPT was evaluated and compared to the original text by blinded reviewers. While ChatGPT can produce high-quality text and summarize information efficiently, its current ability to interpret data and answer scientific questions is limited, and it cannot be relied upon for a literature search or accurate source citation due to the potential spread of incomplete or false information. We advocate for open discussions within the reproductive medicine research community to explore the advantages and disadvantages of implementing this AI technology. Researchers and reviewers should be informed about AI language models, and we encourage authors to transparently disclose their use.
Keywords: ChatGPT; academic writing; artificial intelligence; language models; reproductive medicine.
© The Author(s) 2023. Published by Oxford University Press on behalf of European Society of Human Reproduction and Embryology. All rights reserved. For permissions, please email: journals.permissions@oup.com.