ChatGPT for Automated Writing Evaluation in Scholarly Writing Instruction

Abstract

Background: Effective strategies for developing scholarly writing skills in postsecondary nursing students are needed. Generative artificial intelligence (GAI) tools, such as ChatGPT, for automated writing evaluation (AWE) hold promise for mitigating challenges associated with scholarly writing instruction in nursing education. This article explores the suitability of ChatGPT for AWE in writing instruction. Method: ChatGPT feedback on 42 nursing student texts from the Michigan Corpus of Upper-Level Student Papers was assessed. Assessment criteria were derived from recent AWE research. Results: ChatGPT demonstrated utility as an AWE tool. Its scoring performance demonstrated stricter grading than human raters, related feedback to macro-level writing features, and supported multiple submissions and learner autonomy. Conclusion: Despite concerns surrounding GAI in academia, educators can accelerate the feedback process without increasing their workload, and students can receive individualized feedback by incorporating AWE provided by ChatGPT into the writing process. [ J Nurs Educ . 2023;62(12):721–727.]