Generative artificial intelligence (AI) is a key driver of innovation across all sectors and in education it has the power to optimise teaching and learning to benefit educators and students alike. However, the increasing prominence and influence of AI in domains previously exclusive to humans, such as design, raises urgent questions about the assessment of learning in design education. Recent writings in the field of design education agree that in the age of AI, educators need to revisit existing assessment practices. Conversations about AI and assessment practices appear to revolve around upholding academic integrity and defining what should be assessed when students can create design outcomes using generative AI. This study flips the script: Instead of focusing on managing the use of AI by students to create design outcomes, this study asks a further question. How can design educators use generative AI to facilitate the design of authentic assessment rubrics?

An outcomes-based approach is common in South African design education, and practical projects are used as assessment instruments. The focus of the assessments is on helping students apply their knowledge and skills practically in a real-world context. As a result, design assessments lean towards assessing the design process rather than relying solely on the design outcome, therefore, assessing students’ ability to articulate the reasoning behind their application of knowledge and skills. Assessments used for practical projects typically use assessment rubrics that feature clear evaluation criteria, and aim to provide comprehensive feedback to students. Using authentic assessment, a student-centred approach to assessment practices, curriculum designers can effectively mimic the real working environment, align to the principles and dimensions of authentic assessment, promote deeper learning, and ensure that holistic learning experiences for students are taken into account. However, defining assessment criteria and descriptors that are authentic, clear, well-formulated, and not overly restrictive demands careful consideration to avoid gaps, making it a time-consuming process.

This paper explores the potential of using AI by using ChatGPT in the design of authentic assessment rubrics, and presents a critical reflection on insights gained from action research. The research was conducted by three curriculum designers working across various design qualifications at a private higher education provider. The study contributes to the emerging broader discussion on the innovative use of AI to support teaching and learning within the field of design education.