Artificial Intelligence and the Professional Work of Physics Teachers in STEAM Education

The rapid integration of artificial intelligence (AI) into educational environments is revolutionizing both learning technologies and the professional responsibilities of educators. In physics education, especially in STEAM classrooms, AI is progressively involved in explanation, representation, and problem-solving, transforming conventional distributions of epistemic authority. This Conceptual Paper analyzes the transition of AI from a technical instrument to a cognitive collaborator, focusing on teacher professional expertise. Based on physics education research and recent investigations of AI-assisted instruction, we contend that the educational efficacy of AI is more contingent upon instructors' pedagogical content knowledge, epistemic judgment, and ability to manage learning settings than on its technical complexity. When educators regard AI as a temporary collaborator to be examined through modeling, evidence, and empirical reasoning, AI can enhance inquiry, differentiation, and student autonomy. When regarded as an authoritative source or a shortcut to answers, it jeopardizes sense-making and undermines epistemic responsibility. The analysis emphasizes three interconnected d