Dear Editor,

We found that the article on “A comparison of ChatGPT-generated articles with human-written articles [1]” is interesting. The accuracy and caliber of various academic publications produced by ChatGPT are compared to those authored by human authors by Ariyaratne et al. [1]. According to Ariyaratne et al., ChatGPT can produce coherent research articles that, at first glance, may closely resemble genuine articles written by academic academics and published in scholarly journals [1]. However, Ariyaratne et al. noted that all of the publications we evaluated contained false references and factual errors [1]. However, Ariyaratne et al. came to the conclusion that it is important to note that the papers produced can seem genuine to an uninformed reader [1].

The study’s findings revealed that four of ChatGPT’s five articles were highly erroneous, and all of them had bogus references. One piece, on the other hand, was well written, with a strong introduction and debate section. Even the well-written piece contained false references. These findings imply that, while ChatGPT is capable of producing cohesive and well-structured papers, its correctness and dependability in the context of academic writing, particularly in the field of radiology, are dubious. The prevalence of false references calls into question the legitimacy and veracity of ChatGPT’s material.

It is crucial to highlight that the focus of this study was on the accuracy and quality of radiology articles created by ChatGPT, and the findings may not be applicable to other fields or types of articles. Furthermore, the study did not include information on the exact prompts or instructions supplied to ChatGPT, which could have an impact on the quality and correctness of the generated content. More research is needed to investigate ChatGPT’s limitations and potential enhancements in academic writing. To ensure that the information generated by AI systems like ChatGPT respects academic integrity standards and contributes to the growth of knowledge in relevant domains, challenges like as correctness, fact-checking, and reference must be addressed.

To guarantee that the advantages and hazards of generative AI are balanced, effective governance and monitoring methods must be established. The development, modification, or acceptance of sensitive content by AI should possibly be subject to human review [2]. ChatGPT can offer a wealth of knowledge on problems and suggestions. The results of the ChatGPT suggest that some of these datasets might hold untrue beliefs or assumptions. Patients may therefore receive false or misleading information. It is important to think about any potential ethical issues before using AI chatbots in academic research. It should have been thoroughly probed whether there were any biases in the data or algorithms, authorship attribution, or intellectual property rights.