A systematic review of the evaluation in K-12 artificial intelligence education from 2013 to 2022

This study reviewed evaluations of K-12 AI education from 2013 to 2022. The analysis encompassed 36 articles and focused on examining research methods, sample sizes, evaluation methods and types, learning outcomes, evaluation contexts, and primary findings. The results showed that most evaluations took a summative approach and relied on self-report surveys to evaluate cognitive learning outcomes related to machine learning (ML) concepts among middle and high school students in informal settings. The tools commonly employed for evaluation were AI-based programming tools and online platforms. These findings offer valuable insights for AI educators and researchers in K-12 education, highlighting the significance of introducing students to the assessment of AI in formal learning settings at an early age, utilizing thorough qualitative and mixed-methods techniques, and embracing diverse assessment approaches to evaluate cognitive learning outcomes beyond machine learning.