AI for Academic Writing: Insights from the ASCEND 2026 International Guest Lecture
The Airlangga Global Engagement (AGE) hosted an ASCEND 2026 seminar titled ‘AI for Academic Excellence: Enhancing Scholarly Writing and Publication Readiness through Collaboration with The University of Melbourne’ on Monday, 27 July 2026, via Zoom. The event was part of the ASCEND 2026 programme, a cross-institutional collaboration between Universitas Airlangga and The University of Melbourne aimed at strengthening academic capacity and international networks.
Haeun Kim, PhD, a lecturer in Applied Linguistics at The University of Melbourne, presented her research on the use of generative AI in academic writing. She noted that many ChatGPT users tend to issue short, generic commands such as ‘correct this’ or ‘proofread this’, whereas more detailed prompts yield significantly better and more specific feedback. Her study, which tested seven different command types on academic essays, found that prompts like ‘revise this’ and ‘rewrite this’ triggered more substantial text modifications compared to basic grammar checks.
A second study examined the reliability of AI as an academic evaluator. Using essays from an English Placement Test, the research compared AI-generated scores with human assessments. The findings showed that ChatGPT was moderately reliable for evaluating grammar and vocabulary but was less effective in judging argumentation, content depth, and overall coherence. Dr Kim stressed that AI should be positioned as a complementary tool rather than a replacement for human evaluators, particularly for high-stakes academic assessments.
The presentation also covered the rise of multimodal AI, which can process text, images, audio, and video simultaneously. Dr Kim noted that this advancement creates new opportunities for language learning, especially in speaking, listening, and interactive communication. During the discussion, participants raised concerns about student over-reliance on AI. Dr Kim recommended developing both AI-assisted and independent writing skills. On the impact of generative AI on scholarly publishing, she observed that while AI has democratised access to academic writing for non-native English speakers, it has also led to an increased volume of manuscript submissions, adding to the workload of editors and reviewers and raising quality concerns. She advised exploring various platforms such as ChatGPT, Gemini, Claude, and Llama to find the best fit for individual needs, as no single model excels at all tasks. The session concluded with participants sharing prompting strategies, with many favouring detailed queries over short commands.