SUSTAINABLE AI USE IN EAP: STUDENT PERCEPTIONS AND PRACTICAL APPLICATIONS

In Hong Kong, Generative Artificial Intelligence (GenAI) has been generally well received, with more positive than negative media coverage (Lee & Chen, 2024), favourable student perceptions (Chan & Hu, 2023), and, after an initial period of uncertainty, positive reception within the university community (Cheng & Yim, 2024). However, as GenAI tools advance, concerns have emerged regarding teachers'
overconfidence in their ability to detect AI-generated texts (Fleckenstein et al., 2024). Studies also indicate that students' over-reliance on GenAI tool may lead to over-dependence and lack of student agency (Darvishi et al., 2024) and poorer learning outcomes (Glass & Kang, 2020). These issues raise reasonable concerns about GenAI’s short-term and long-term impacts on English for Academic Purposes (EAP)
teaching and learning practices.
This study investigates the attitudes towards GenAI and the experiences with AI-driven tools of students (n=76) from four English Language Centre subjects at a Hong Kong university. The research questions addressed include: What are the students' perceptions of GenAI in their learning process? How do AI-driven tools influence their academic performance and learning outcomes? How can EAP practitioners develop
balanced teaching strategies that utilise the affordances of GenAI while maintaining academic integrity and fostering genuine learning? To answer these questions, the researchers employed a mixed-methods approach, collecting written assessments, GenAI transcripts, and student reflections. Additionally, pre- and post-course surveys and interviews were conducted to gather comprehensive data on student
experiences and attitudes.
The findings of this study highlight both the challenges and benefits of integrating GenAI into EAP courses, underscoring the importance of developing practical and sustainable applications, strategies, and assessments to counterbalance students’ use of GenAI tools. Writing samples and student responses raise concerns about negative consequences of GenAI use such as hindering the development of critical
thinking and creative problem-solving skills; decreasing reading skills and time spent reading; increasing the inaccurate use of low-frequency vocabulary items and inappropriate expressions; undermining academic integrity and ethics standards; de-emphasizing the importance of peer collaboration and human-to-human interactions during the learning process; and circumventing productive struggle. However, the data also suggests GenAI tools have the potential to increase student engagement through an adaptive and interactive approach to texts; assist in preliminary brainstorming and research stages; and enhance academic English skills by providing immediate and personalised feedback on content and language production and for a wider range of learning preferences and L2 proficiency levels.
In this presentation, we will elaborate on these findings, explore specific student examples, and discuss their implications regarding teaching strategies that leverage GenAI while still supporting the achievement of authentic language proficiency and other learning outcomes.