Integrating Artificial Intelligence in STEM Classrooms: A Mixed-Methods Study of Models, Outcomes, and Ethical Challenges

This paper investigates the integration of Artificial Intelligence (AI) and Machine Learning (ML) into STEM (Science, Technology, Engineering, Mathematics) education using a mixed-methods approach. It proposes a Strategic AI Integration Framework encompassing curriculum redesign, intelligent tutoring systems, adaptive learning, and virtual laboratories. The study combines a literature review, case analyses, and a validated survey of 154 teachers and 312 students across K-12 and higher education settings. Pilot testing yielded a Cronbach's alpha of 0.84, confirming instrument reliability. Experimental results show improved engagement among the students. Also providing conceptual clarity and instructional support. The framework discussed in the study uses AI techniques such as Bayesian Knowledge Tracing, Q-learning, and BERT, which are integrated with real classroom simulations. Various issues, such as ethical concerns, algorithmic bias, and data privacy, are also discussed in the study. The research findings contribute actionable insights for educators, policymakers, and technology developers aiming to promote an inclusive, scalable, and ethically guided AI integration in STEM education.