Transforming Academic English Pedagogy: GenAI-Integrated Tasks, AI Literacy, and the Psychology of Student Engagement

International Summit on the Use of AI in Language Learning and Teaching 2026 (AIinLT 2026, 22 June).

​The Hong Kong Polytechnic University for the International Summit on the Use of AI in Language Learning and Teaching (AIinLT 2026) took place from 21–24 June 2026. Organised by the Department of English and Communication in partnership with the Education Bureau and SCOLAR, the summit explored the theme “AI in Language Education: From Ideas to Implementation to Impact.”
On 22 June, our team presented Transforming Academic English Pedagogy: GenAI-Integrated Tasks, AI Literacy, and the Psychology of Student Engagement.
The mixed-methods study examined how GenAI-integrated tasks, paired with AI literacy instruction, addressed students’ psychological needs for relatedness, competence, and autonomy (Self-Determination Theory) in first-year English for Academic Purposes and content-based courses across six Hong Kong universities. We investigated impacts on AI literacy (usage/competence, critical evaluation, ethical awareness) and cognitive, behavioural, and social engagement.
Highlights: Quantitative data from 249 paired pre- and post-questionnaires showed significant gains in all three AI literacy dimensions (small effect sizes). Students also reported high perceived social, cognitive, and behavioural engagement. Multiple regressions revealed that post-intervention critical evaluation strongly predicted all three forms of engagement; ethical awareness further predicted cognitive and behavioural engagement.
Qualitative interviews with 23 students highlighted six themes: enhanced AI literacy (prompt crafting, metacognitive strategies, critical debate with AI); mixed views on academic English competence (helpful for writing, less so for speaking); growth in autonomy and self-efficacy alongside risks of over-reliance; stronger cognitive-behavioural engagement through well-sequenced tasks, tempered by occasional disengagement; mixed social outcomes (better peer communication versus potential isolation); and a clear preference for human interaction and limited trust in AI feedback.
Takeaways: AI-integrated tasks work best when they deliberately foster social interaction and bonding, prioritise face-to-face connection and empathy, and embed robust AI literacy. Future work would benefit from control groups, more speaking-focused activities, performance measures, and longitudinal designs.