Work

Authors’ Reply: Critical Limitations in Comparing ChatGPT and DeepSeek for Orthopedic Assessment

Year: 2026

Type: article

Abstract: We respond to comments on our study comparing ChatGPT and DeepSeek for answering orthopedic multiple-choice questions. We clarify that the reported Cohen κ values reflect inter-rater reliability within each model rather than agreement between the two models. All questions were administered in English, and the findings therefore reflect performance in an English-language context. We acknowledge limitations related to reproducibility due to the use of web-based interfaces and address concerns about data contamination. We also correct a typographical error in the reported accuracy for the pelvic and spine injury category.

Source: JMIR Formative Research

DOI: 10.2196/91470

Authors: Chirathit Anusitviwat, Sitthiphong Suwannaphisit, Jongdee Bvonpanttarananon, Boonsin Tangtrakulwanich

Institutions: Prince of Songkla University, Navamindradhiraj University

Language: English

Topic: Artificial Intelligence in Healthcare and Education

Subfield: Health Informatics

Field: Medicine

Domain: Health Sciences

SDG: Good health and well-being

Open Access status: gold

Estimated APC paid: $1,900