Investigating AI Imitative Capabilities: A Comparative Genre Analysis of Reddit and ChatGPT-4 Spouse Appreciation Posts
ChatGPT-4 has been a helping tool for its users in various fields and contexts. Among these, a recurring topic of inquiry is AI’s ability to mimic human language, particularly appreciative sentiments. In this study, we explore the current capabilities of ChatGPT-4 at mimicking human appreciatory language by comparing its prompt-based generated texts to the appreciative posts of married individuals on Reddit, specifically subreddit r/Marriage. Using Swales' moves analysis as the working framework, we determined and then labelled moves found in the Reddit posts as we examined them as a genre and compared them with AI-generated posts. Findings revealed that ChatGPT-4 imitated 11 out of the 13 moves present in the Reddit posts. Furthermore, five categories of obligatory moves were established to better distinguish between the texts. ChatGPT-4 was observed to have more clearly obligatory moves and was more consistent in following a certain sequence than texts written by a human being. Additionally, ChatGPT-4 generated texts had a higher average word count and gave an impersonal impression due to its lack of the following moves: Description of Spouse, Description of Original Poster, and Relationship Background. With these findings, we recommend further linguistic studies on Reddit posts and other non-academic social media forms as a genre and the continuous use of genre analysis to circumscribe, more precisely, the AI imitative capabilities of human natural language and its implications to future applications. Keywords: artificial intelligence; ChatGPT-4; imitative capabilities; Reddit; moves analysis
Authors
- J. M. De Miranda
- Marielle Jessie Dela Peña De Guzman
- Erika Kristine Estanislao Arcenal
- Licca Pauleen Valdez Capistrano
- Micaela Isabel Molina Forrosuelo
Institutions
- La Salle University (US)
- De La Salle University (PH)
- University of the Philippines Diliman (PH)
Publication Details
- Journal
- 3L The Southeast Asian Journal of English Language Studies
- Published
- 2025-03-26
- DOI
- https://doi.org/10.17576/3l-2025-3101-04
- Primary Topic
- Artificial Intelligence in Healthcare and Education
- Type
- article
- Field-Weighted Citation Impact
- 0.00