Adoption and pedagogical use of generative artificial intelligence in K-12 mathematics/science/language classrooms: Challenges and opportunities

Abstract

Generative artificial intelligence in K-12 mathematics, science, and language classrooms has already presented major opportunities of personalized learning, automatically provided feedback, curriculum development, and student interaction, and posed multiple questions about ethical AI, academic dishonesty, teacher preparedness, algorithm bias, and student data security. With the prevalence of large language models, generative AI-driven learning aids and multimodal learning platforms that have begun to be popular in schools, more individuals are interested in learning about the pedagogical utilization, utility, limitations, and implications of Generative AI in school learning. This literature review is based on the PRISMA framework that provided a systematic inspection of the current research of the adoption and pedagogical integration of Generative AI in K-12 mathematics education, science education, and language education. The review conducted to synthesize the evidence on the identification of the dominant themes in research, patterns of implementation, and gaps in current scholarship. Special focus was put on adaptive learning, intelligent tutoring systems, automated assessment, teacher professional development, multilingual learning, differentiated instruction and human-AI collaboration. The results demonstrate that personalized learning, cognitive scaffolding, inquiry-based learning, project-based learning, formative assessment, and self-regulated learning can be greatly enhanced in mathematics, science, and language classrooms with the help of Generative AI. It is concluded that responsible AI systems, enhanced teacher education, fair access and evidence-based pedagogical models balancing the innovativeness with human regulation are the needed components of future-ready deployment of Generative AI in K-12 education.