A Multimodal Interactive Framework for Science Assessment in the Era of Generative Artificial Intelligence (Poster 11)

• In Event: AERA Poster Session Sunday 9:50 a.m.
  In Poster Session: Emerging Issues in Classroom Assessment

Abstract

The rapid evolution of Artificial Intelligence (AI) is shaping education while also raising concerns. Of particular concern is the growing use of generative AI tools to outsource assessment tasks, compromising academic integrity and student learning. Addressing these challenges necessitates reflections on existent assessment tasks. This study advances a framework to analyze tasks through the lens of multimodality and interactivity. We present the framework through detailed descriptions of four categories along with concrete examples in each category. We analyzed how prompt and response affordances of each category allow to assess science constructs, while also rendering them resistant to AI-based outsourcing. We conclude by discussing various ways in which the framework can serve as an analytical tool for educational researchers and practitioners.

Authors

• Yizhu Gao, University of Georgia

• Xiaoming Zhai, University of Georgia

• Min Li, University of Washington

• Xiaoxiao Liu, University of Alberta

• Ying Cui, University of Alberta