Abstract
Background/context
Recent discussions about GenAI in education has primarily emphasized developing (Gen)AI literacy among students, including the ethical and effective use of GenAI for learning. However, less attention has been given to the potential application of GenAI as a tool to support teachers’ work and productivity. Research has suggested that integrating GenAI programs can facilitate teachers’ idea generation, materials creation, grading, and feedback for students, thereby freeing valuable time for teachers to focus on instructional quality and deeper pedagogical engagement (Law, 2024). Yet, teachers’ concerns regarding the accuracy, reliability, and trustworthiness of GenAI-generated output remain prominent. Educators must evaluate GenAI-produced examples with a critical lens, ensuring the linguistic, factual, and contextual appropriateness of materials before introducing them to students. An even more pressing concern involves the security and privacy of student data. The process of uploading students’ work or classroom materials to online GenAI program servers via the internet raises ethical and security questions about the implications of sharing students’ intellectual property, personal information, and potentially sensitive data with third-party services whose data handling practices may be opaque or insufficiently regulated. Addressing these risks is essential for building teacher confidence and for upholding the ethical standards of educational institutions.
Description of initiative or practice
Drawing on my own interdisciplinary exploration and practical experience in creating materials for several CAES courses using various GenAI tools, this presentation will share some insights from using secure GenAI tools to assist materials creation as well as some essential steps that can help minimise the risks of data leakage. I will outline the use of the HKU ChatGPT platform, which offers enhanced privacy protections tailored for educational settings. Furthermore, I will discuss the implementation of offline GenAI solutions where data does not leave the personal computer, drastically reducing exposure to external risks. Through real-world classroom scenarios, I will demonstrate how secure GenAI can facilitate the creation of high-quality, contextually relevant materials while safeguarding student privacy and institutional integrity.
Evidence of outcomes/effectiveness
Effort in creating these materials is much appreciated by 2025–2026 semester 2 CAES9532 students, as reflected in the SFTL mean score of 100 for overall course effectiveness (overall response rate: 73%). Notable student comments include “Energetic, so much effort on every class, and great course materials”, and “Downright the most enthusiastic and energetic course. Variety of course materials and the effort of the instructor makes this course enjoyable”.
Presenter(s) / Author(s)
