AI-Driven Learning Analytics for Evidence-Based Decision-Making in K–12 STEM Classrooms

Abstract

Artificial Intelligence (AI) and Learning Analytics (LA) are redefining personalized learning space and instruction with real-time feedback and data driven pedagogy in K-12 STEM education. The present study aims to document and assess the use of Al and LA tools in public in Lahore, Pakistan, over five years. The study employing mixed methods incorporated classroom observation and interviews with teachers and student's performance data to assess the value and the cost of the technology. Results indicate that Al and LA tools contribute to better engagement and performance of students in STEM disciplines, especially in Mathematics, Science, and Physics. On the other hand, Poor infrastructure, absence of pre-service and in service teacher training, and the digital divide constrain technology driven education to a few. The study presents a data driven learning design informed by the concepts of a technology empowered learning community, instructional design and Al to address the issues of pedagogy to enhance students learning outcomes. The work substantiates the need of focused policy efforts and systemic professional development in underfunded schools to meet the fundamental right of students to learn by pedagogy that harness the power of AI.