From Proposal to PhD: Building an HDR Pipeline Through AI-Integrated Assessment

Abstract

Traditional groupwork in higher education is plagued by well-documented challenges: free-riding, scheduling conflicts, small class sizes, and superficial collaboration. Simultaneously, the rapid emergence of generative AI demands that educators move beyond prohibition towards meaningful integration. This presentation describes the design and first-iteration outcomes of a cross-disciplinary capstone assessment (SCI310/510 – Artificial Intelligence and Computing for Science) in which students assemble “AI teams” – multiple AI assistants configured as discipline-specific experts – to collaboratively develop research proposals addressing real-world grand challenges. Students operate as project leads, critically evaluating AI outputs, documenting prompt refinement, and validating claims against primary literature. Authentic understanding is verified through a mandatory viva voce, ensuring academic integrity without restricting AI use during the creative process. Critically, the assessment is deliberately scaffolded to mirror the early stages of a Higher Degree Research candidature: students identify research gaps, write formal proposals, develop budgets and timelines, and defend their ideas under questioning – functioning as an “HDR pipeline” that demystifies the pathway to Honours and PhD study. First-offering evaluation data (4.7/5.0) and student feedback indicate high engagement, with students reporting the experience provided practical insight into academic research life. Implications for assessment design, AI literacy, and HDR recruitment are discussed.