Conference Agenda

Generative artificial intelligence (GenAI) is rapidly transforming higher education. However, there is limited quantitative data on how its use affects students' experiences with assessments and learning support. This study explores whether different levels of GenAI usage are linked to students' perceptions of assessment design quality and learning support in Australian higher education. Using a cross-sectional quantitative survey design, the study initially got 107 responses. After screening it retained 68 complete cases for the main analysis. Three composite indices were examined: GenAI use, perceived learning support, and perceived assessment design quality. Descriptive statistics were followed by one-way analysis of variance (ANOVA) comparing low, moderate, and high GenAI-use groups. The result shows a statistically significant difference in perceived assessment design quality across GenAI use groups. F(2, 59) = 3.464, p = .038. Perceived learning support did not differ significantly, F(2, 65) = 2.832, p = .066. These findings indicate that the most noticeable impact of GenAI in this dataset lies in assessment redesign rather than in broader changes to students’ day-to-day support experience. By providing empirical data from the Australian higher education context, the paper builds on the ongoing conversation about responsible AI integration. It also emphasises the need to translate institutional AI policy into specific assessment practices that preserve educational quality and uphold pedagogical integrity.

Generative artificial intelligence (GenAI) is rapidly becoming adopted in academic research practices, particularly among STEM graduate students. While much attention has been given to the technical capabilities of GenAI tools, far less is known about how graduate students conceptualize their role in research when working with such technologies and how theoretical frameworks shape these conceptions. This study examines STEM graduate students’ perceptions of the relevance of Diffusion of Innovations theory to their academic work in the GenAI era. Specifically, we show that they perceive themselves, as well as their peers and study environment, as being very innovative and discuss the implication of this finding in general and in the age of GenAI in particular.

This work presents an automated system for generating Systematic Literature Reviews (SLR) by integrating the Gemini 2.5 Flash Large Language Model with a refined multi-agent architecture and iterative refinement cycles. The proposed solution aims to streamline the SLR process, encompassing literature collection, analysis, and final document generation.
Key features include automated paper retrieval via the Arxiv and Scopus APIs, comprehensive support for manual PDF uploads, and the dynamic creation of visual analytics (charts and graphs). By leveraging a specialized multi-agent system, where each agent’s output informs subsequent agents, the approach aims to enhance content consistency and mitigate hallucinations.
The system further incorporates a human-like writing style emulation, drawing inspiration from provided academic texts to produce natural and professional prose. The final output is a fully formatted LATEXdocument ready for academic publication, potentially requiring minimal human input and contributing to research synthesis. This contribution addresses the growing demand for automated research synthesis in academia, offering researchers an efficient tool for producing systematic literature reviews.