Nick Wilson

Academic misconduct represents an assessment design challenge, rather than a security challenge. We cannot secure everything, and technology will always be able to adapt to any security measures put in place. Drawing upon de-identified cases from Academic Integrity Matters, this interactive workshop provides examples of how concerns were first spotted, what further investigations uncovered, and how assessment redesign could have reduced the likelihood of the misconduct occurring.
Participants will explore the relationship between assessment design, assurance of learning, and academic integrity through a series of practical case studies. The workshop will then shift perspective, inviting participants to step into the role of an Academic Integrity Investigator. Working in small groups, we will analyse a fictitious misconduct case using evidence similar to that considered during a formal investigation. Participants will evaluate the available evidence, determine appropriate next steps, consider a student response, and apply a penalty determination matrix to reach an outcome. This will prompt participants to reflect on the challenges faced by academic integrity investigators when determining outcomes in misconduct cases.
The session will provide practical insights into designing assessments that promote authentic learning while reducing opportunities for academic misconduct.

Generative artificial intelligence (GenAI) and Tertiary Education Quality Standards Agency (TEQSA) requirements are catalysts for driving higher education institutions to reflect on and uplift assessment products so that they preserve academic integrity and evidence student outcomes. This presentation will analyse quantitative data extracted from the Curriculum Information Repository (CIR) (Akari Software). More specifically, the number and diversity of assessment products across the University of the Sunshine Coast, including the most prevalent and least frequently used assessment types. Additionally, we explore the distribution of assessment products across the five Schools. Most importantly we showcase opportunities and practical strategies for uplifting assessment products to enhance assured learning for both blended and asynchronous delivery modes.

The natural process of understanding speech involves combining auditory and visual cues. CNN based lip reading systems have become very popular in recent years. However, many of these systems consider lipreading to be a black box problem, with limited detailed performance analysis. In this paper, we performed transfer learning by training the Inception v3 CNN model, which has pre-trained weights produced from IMAGENET, with the GRID corpus, delivering good speech recognition results, with 0.61 precision, 0.53 recall, and 0.51 F1-score. The lip reading model was able to automatically learn pertinent features, demonstrated using visualisation, and achieve speaker-independent results comparable to human lip readers on the GRID corpus. We also identify limitations that match those of humans, therefore limiting potential deep learning performance in real world situations.