From Prompt to Practice: Using AI to Build Better Assessment Rubrics with SOLO Taxonomy

Designing effective assessment rubrics in higher education is a time-intensive task that demands clarity, fairness, and differentiation. This paper explores the use of generative artificial intelligence, specifically ChatGPT-4, to streamline the rubric development process, with a focus on alignment to the SOLO Taxonomy. SOLO (Structure of the Observed Learning Outcome) offers a hierarchical model for describing student performance, ranging from prestructural to extended abstract levels. Integrating AI into rubric design helped address key challenges such as consistency, objectivity, and reducing workload, while maintaining academic oversight and alignment to institutional policies on responsible AI use.

The author outlines a practical approach using AI prompts that incorporate assessment criteria and SOLO-based banding to generate structured performance descriptors. While AI reduced design time and enhanced feedback clarity, limitations such as overgeneralised language and the risk of inaccuracies highlight the need for critical human review. Audio feedback was employed to personalise the otherwise standardised rubric outputs.

Reflections underscore the evolving role of AI in assessment practices and advocate for further exploration into student experiences with AI-informed feedback. The author encourages collaborative development and sharing of AI-enhanced practices within academic communities to ensure quality and innovation. While AI is not a replacement for academic judgment, it is a powerful tool for supporting efficient and pedagogically sound assessment design.