Sarcouncil Journal of Education and Sociology

Abstract: Artificial intelligence (AI) is increasingly transforming educational practices, including the teaching and learning of physics. However, the educational significance of AI depends not only on its technological capabilities but also on how effectively it collaborates with students and teachers within pedagogically meaningful learning environments. This systematic literature review examines the development of human–AI collaboration in physics learning and identifies emerging pedagogical models reported in the literature. Following the PRISMA 2020 framework, relevant studies were systematically identified, screened, assessed for eligibility, and synthesized according to their AI applications, human roles, physics learning activities, learning outcomes, and reported challenges. The synthesis identified four major pedagogical models: AI as an intelligent tutor, AI as a cognitive partner, AI as a feedback assistant, and teacher-mediated human–AI collaboration. The findings indicate that AI can support conceptual understanding, physics problem solving, critical thinking, self-regulated learning, scientific reasoning, and learner engagement. However, the educational benefits of AI are strongly influenced by the degree of student participation and teacher mediation. The review also identifies persistent challenges involving inaccurate AI-generated content, student overreliance, limited AI literacy, assessment integrity, teacher preparedness, privacy, and unequal access. Overall, the findings support a human-centered approach in which AI augments rather than replaces human reasoning. Students should remain active evaluators and decision-makers, while teachers retain responsibility for pedagogical design, disciplinary validation, and responsible AI use. The review provides a conceptual basis for developing emerging human–AI pedagogical models that promote meaningful, critical, and responsible physics learning.