Abstract This study presents the first comparative analysis of appraisal patterns in academic book reviews generated by ChatGPT and those authored by humans. Utilizing the Appraisal Framework, we identify distinct evaluative profiles across three subsystems: Attitude, Engagement, and Graduation. Findings indicate that while both artificial intelligence and human authors primarily employ Appreciation resources, significant differences exist in their use of Affect and Judgment, with human-authored reviews showing a richer and more nuanced expression of emotion and evaluation. Human writers also demonstrate greater flexibility in employing Engagement strategies and Graduation resources, fostering a more dynamic reader relationship. Conversely, ChatGPT-generated reviews, though structurally coherent, reveal a limited capacity for skilled interpersonal Engagement, resulting in a more impersonal and less persuasive evaluative stance. These insights underscore the limitations of current large language models in replicating the rhetorical depth of human writing, highlighting implications for English writing pedagogy.
Yao et al. (Tue,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: