Speaker
Description
Generative artificial intelligence (GenAI) is increasingly used in university students’ academic research, yet responsible use requires more than access to AI tools or prompt-writing skills. Students also need to critically evaluate AI-generated information, verify sources, disclose AI use, revise AI outputs, and remain accountable for final academic products. This study describes the design and preliminary evaluation of a multi-component training model intended to operationalise responsible GenAI use in academic research within teacher education.
The model integrated foundational GenAI instruction, prompt-writing training, AI tools for research support, academic ethics, a research-based post-training assignment, an AI Transparency Appendix, rubric-based assessment, and post-training support. A descriptive preliminary evaluation was conducted with 23 student research groups comprising 79 teacher education students in Viet Nam. Post-training products were assessed using a 100-point rubric covering research product quality and responsible AI use, supplemented by a post-training survey and open-ended responses.
Of the 23 groups, 11 submitted complete research assignments, nine submitted partial AI-related evidence, and three provided no assessable submission. Among complete submissions, the mean total score was 63.82 ± 11.42. Students performed relatively better in methodology and prompt quality/tool appropriateness, but showed weaker performance in literature review, AI transparency, source verification, and critical revision of AI-generated outputs. Survey responses indicated positive perceptions of AI-related knowledge, ethics awareness, prompting competence, and training usefulness, while reported frequent AI-use behaviours were lower.
The findings suggest that responsible GenAI education should move beyond tool introduction toward structured, product-based, and transparency-oriented learning designs. Integrating transparency documentation with rubric-based assessment may help make responsible AI practices more observable, assessable, and supportable in student academic research.