Portal de Periódicos da CAPES

Promoting Agency Among Upper Elementary School Teachers and Students with an Artificial Intelligence Machine Learning System to Score Performance-Based Science Assessments

2025; Multidisciplinary Digital Publishing Institute; Volume: 15; Issue: 1 Linguagem: Inglês

10.3390/educsci15010054

2227-7102

Fatima E. Terrazas-Arellanes, Lisa A. Strycker, Giani Gabriel Alvez, Brian G. Miller, Kathryn Vargas,

Educational Games and Gamification

As schools increasingly adopt multidimensional, phenomenon-based, digital-technology-enhanced science instruction, a concurrent shift is occurring in student performance assessment. Assessment instruments capable of measuring multiple dimensions must incorporate constructed responses to probe students’ ability to explain scientific phenomena and solve problems. Such assessments, unlike traditional multiple-choice tests, are time-consuming and labor-intensive for teachers to score. This study investigates the potential of an artificial intelligence machine learning system (AI-MLS) to address two critical questions: (1) How accurately can the AI-MLS replicate human scoring of multidimensional science assessments? and (2) How can the implementation of AI-MLS promote educational equity and reduce teacher workload? The present paper describes the development of the AI-MLS to rapidly and accurately score third- to fifth-grade students’ constructed responses on multidimensional science assessments. It summarizes key findings from the study, discusses findings in the broader context of fostering agency through digital technology, and offers insights into how artificial intelligence technology can be harnessed to support independent action and decision-making by teachers and students.