Annishah (2025) Investigating efl students’ generative artificial intelligence (ai) literacy in academic writing / Annishah</p>. Masters thesis, Universitas Negeri Malang.
Full text not available from this repository.Abstract
Writing in the academic field is often seen as a demanding task. This triggers the use of technological tools such as generative AI to assist them in writing. One of the most used generative AI tools is ChatGPT. The existing studies show that ChatGPT has positive and negative sides for education especially when used during writing. ChatGPT was believed to have the ability to help students brainstorm ideas act as a private tutor and even be a stress reliever. On the other hand ChatGPT posed some threats such as increasing memory loss among students decreasing academic skills overreliance and challenges related to ethics and data bias. To prevent these negative effects of ChatGPT utilization during academic writing AI literacy especially generative AI literacy is important to be discussed as it is believed to promote responsible use of AI including ChatGPT. Therefore this study aims to (1) investigate perceived generative AI literacy among EFL students (2) reveal the utilization of ChatGPT in academic writing across generative AI literacy levels and (3) uncover the possible factors causing various levels of generative AI literacy. The methods used in this study are explanatory mixed methods combining quantitative and qualitative methods. The participants were undergraduate EFL students in a state university in East Java who were selected based on purposive sampling. Two hundred seven students participated in the quantitative phase. They were from batch 2020- 2023. The second phase qualitative involved ten participants who joined semi-structured interviews. Two kinds of instruments were used to collect data which were a questionnaire and a list of interview questions. The survey was conducted online using Google Forms. The items were adapted from Ng et al. (2023) Ng et al. (2024) Alzubi (2024) and Wang et al. (2022). Before the data collection some experts were asked to validate the items. Finally there were 32 items for the survey and 10 questions for the semi-structured interview. The interviews were all conducted online. They were recorded and I took some important notes during interviews. After collecting survey data the quantitative data was analyzed using SPSS to ensure their validity and reliability and to find the means. The qualitative data were transcribed carefully and analyzed using thematic analysis. The results show that the overall generative AI literacy level of the participants is medium (M 3). Each domain has a different mean. The highest mean belongs to the ethical domain while the lowest one is the behavioral domain. Forty-four students were confident enough to state that they had a high level of generative AI literacy. One hundred forty-seven students belong to the medium level and seventeen students have low levels. The data also suggests that each level of generative AI literacy uses ChatGPT differently in terms of prompting and writing processes. All levels use ChatGPT in brainstorming and outlining. In addition low and medium levels use ChatGPT to generate essays and find references. While the high level utilizes the tool in summarizing and editing which suggests more reflective and responsible practices are done by high levels. The results also suggest that four factors cause different levels of generative AI literacy educational background self-regulation self-efficacy and age. This study has added to the body of knowledge about generative AI literacy in education especially in academic writing among EFL students. The framework of generative AI literacy is also proposed in this study which can be useful for future research. In addition the real practice of ChatGPT utilization is presented. This can act as one of the bases for developing guidance on how to use it properly in academic writing which is needed by students to encourage the appropriate use of ChatGPT. Some limitations are also addressed in this study. First the participants in this study are only EFL students. Further study may include more diverse majors as the results suggest that the background of education leads to different generative AI literacy levels. Second the instrument in the survey was in the form of a self-report. It is suggested to combine self-report and performance-based tests to provide a more complete depiction of generative AI literacy.
| Item Type: | Thesis (Masters) |
|---|---|
| Divisions: | Fakultas Sastra (FS) > Departemen Sastra Inggris (ING) > S2 Pendidikan Bahasa Inggris |
| Depositing User: | library UM |
| Date Deposited: | 04 Jun 2025 04:29 |
| Last Modified: | 09 Sep 2025 03:00 |
| URI: | http://repository.um.ac.id/id/eprint/390592 |
Actions (login required)
 |
View Item |