Future Directions in Classroom Creativity Assessment: Gen AI and Fitness-for-Purpose

This chapter explores the potential of Generative AI to provide fit-for-purpose assessments of product creativity in a classroom context.The chapter begins with a discussion of the drivers behind the need for classroom creativity assessment and focuses on the factors central to the fitness-for-purpose of creativity assessments, including quality, logistics, and authenticity. With a clear picture of the factors that affect fitness-for-purpose set out, the chapter continues by exploring these issues through a custom GPT in the OpenAI/ChatGPT Large Language Model (LLM) environment. The creation of a custom GPT involves configuring the underlying LLM with a set of instructions and a rubric in order to change how the model behaves. In this case, the core of the customisation involved uploading the Creative Solution Diagnosis Scale (CSDS: Cropley, Kaufman & Cropley, 2011) to the model, along with a description of how to score artefacts on this scale, and instructions to provide formative feedback on how different product criteria could be improved.The resulting custom GPT is then tested, first to explore its ability to provide meaningful summative evaluations (i.e., numerical scores) of creativity (by comparing the custom GPT scores with human evaluations of creative writing and mousetrap designs) and second to explore its ability to provide meaningful formative feedback (by examining its evaluation of different paper airplane designs).The chapter concludes by suggesting that generative AI (specifically LLMs) has good potential to address long-standing issues of the fitness-for-purpose of classroom creativity assessments. In the era of Industry 4, and the rise of creativity as an essential 21st century skill, the ability to assess creativity in the classroom is the key to ensuring that students are prepared for the challenges of the future of work.