Phenomenology of AI and ChatGPT Proliferation
Large language models (LLMs), generate significant amounts of new text based on brief input requests. These tools use artificial intelligence to assist in the generation of original content. In contrast with automated paraphrasing tools and digital word assistants that focus primarily on already written text and suggest text changes, these models are trained on broad data sets (Perkins, 2023). ChatGPT is an example of an advanced LLM, which can assist students in developing full-length assignments, with the possibility of including references. Generative pre-trained transformers are fast evolving technologies, and as OpenAI continues to work on more refined versions, this has opened several debates on the impact of ChatGPT on academic integrity (Hern, 2022; Marche, 2022).
When it comes to non-native English students, AI tools can be used simply to reword, reorganize, or paraphrase an already written assignment to the level of the user’s satisfaction prior to submission. The generative power of ChatGPT can paraphrase contents multiple times by adding functionalities or commands requested by the user to improve the quality of the output (Kundu & Bej, 2021). This aligns with the technology acceptance model introduced by Davis in 1985, which suggests that usage of information technology is often influenced by the beliefs the user holds about its perceived usefulness (PU) (Karahanna & Straub, 1999). Based on the PU theory, a person would adopt LLM only if the tool would enhance his/her performance (Korzynski et al., 2023). Accordingly, it is worth noting that some students even go to the extent of paying for additional features from ChatGPT, which contributes to enhancing the overall quality of assessment and user experience. Hence, the perceived usefulness of ChatGPT and AI tools has gained momentum in academia (Gill & Kaur, 2023).
Moreover, when it comes to non-native speakers (NNS), a significant increase in the use of technology and AI, to compensate for linguistic weaknesses, has been observed (Gayed et al., 2022; Zhao, 2023). The intention may not be to cheat, yet the continuous evolution in the world of AI means that there can be overreliance on such tools at the expense of a student learning to improve their academic writing abilities. Automated paraphrasing tool (APT) is yet another example of an application that supports English as a foreign language writer in expressing their ideas more proficiently in written English (Nazari, Shabbir, & Setiawan, 2021). Previous research by Ku et al. (2014) highlighted that some higher-level cognitive and emotive skills are extremely challenging to evaluate using conventional selected answer forms, which therefore, necessitates the adoption of e-assessments as an alternative (Korsgaard, 2020; McKenzie,1998; Pepple, 2022). This, in turn, opens new windows for English as a foreign language writer who can produce better quality assessments using the synergistic effect stemming from AI tools.