AI in higher education

Guidelines on assessment design from Australian universities

Abstract

With the advent of generative AI (GenAI) technologies, academic integrity has become a pressing concern for higher education institutions. This study explored the publicly available guidelines from Australia's Group of Eight (Go8) universities regarding assessment design in response to GenAI tools. The findings indicate a common concern about the challenges GenAI tools pose to traditional assessments, with current AI detection tools deemed insufficient. However, GenAI also presents opportunities to reassess and redesign assessments to enhance academic integrity and learning. The guidelines were provided around two main pieces of advice: exploiting GenAI tools' limitations to make it difficult to complete assessments using GenAI tools and making assessments more relevant to students to increase their engagement and reduce reliance on GenAI. Common recommendations in the guidelines included testing assessments' vulnerability to GenAI, emphasising critical thinking, incorporating contextual elements, designing authentic assessments, using alternative assessment formats, focusing on process-oriented and staged assessments, and using collaborative or in-class assessments. Potential challenges of redesigning assessments as well as limitations of recommended approaches in making assessments totally immune to GenAI tools were also acknowledged. However, ways of addressing these challenges particularly in specific educational contexts remained largely unaddressed. Despite shared concerns and strategies provided in the guidelines to design more rigorous assessments less vulnerable to GenAI tools, the specificity of guidelines varied across institutions. The findings underscore the need for continuous evaluation and adaptation of assessment practices to uphold academic integrity in the face of evolving GenAI technologies.