Designing Automated Writing Evaluation Systems for Ambitious Instruction and Classroom Integration
Published in: Artificial Intelligence in STEM Education: The Paradigmatic Shifts in Research, Education, and Technology, Chapter 13, pages 195–208 (2022). doi: 10.1201/9781003181187
Posted on RAND.org on January 04, 2023
Despite their potential to increase students' learning opportunities in large numbers of schools, tools leveraging artificial intelligence (AI) have had limited uptake by teachers. This is, in part, because technology development and implementation has not always been responsive to the education context in which schools are situated. In the first part of this chapter, the authors draw on sociocultural theory and their research developing an automated writing evaluation system (eRevise) to support argument writing in the intermediate grades to present considerations for the design of AI-based formative feedback tools. They argue that AWE systems should communicate the features of authentic tasks, provide information that is transparent, actionable, and fair, and open up avenues for student-centered classroom discussions and collaboration. The second section of the chapter is informed by education policy implementation research. The authors argue that to support classroom integration and widespread adoption of AWE systems, developers must consider the contextual and subjective factors that influence classroom routines. These factors include the policy context in which classrooms are situated, as well as the values and goals held by school leaders and teachers within that educational system.