Exploring the frontiers of generative AI in assessment: Is there potential for a human‐AI partnership?

Abstract

Abstract This chapter investigates the integration of generative AI (GenAI), specifically ChatGPT, into institutional and course‐level assessment at Western New England University. It explores the potential of GenAI to streamline the assessment process, making it more efficient, equitable, and objective. Through the development of a proprietary GenAI tool, the study examines GenAI's assessment of student evidence, including written assignments and computer coding tasks, against human assessment. It addresses challenges such as data collection, coordination, and the need for well‐defined and precise rubrics. We found notable differences in GenAI and human scoring in some cases, indicating GenAI's potential in certain assessment contexts while also acknowledging its limitations in others. Our research suggests that GenAI could enhance educational assessment processes, but its integration requires addressing biases in training data and securing buy‐in from various stakeholders. Using GenAI to handle particular routine tasks can potentially free up faculty to engage in richer discussions and educational improvements.

Ähnliche Arbeiten