From Prompt to Embodied Simulation: Using Generative AI to Create AR Physics Learning Tools

Synthesis: Levy et al. (2026) show how a structured natural-language prompt can generate a browser-based, hand-controlled augmented-reality (AR) physics simulation — spread your thumb and index finger and a virtual lamp changes color — and describe its use in an introductory physics class. Computer simulations have a long record of supporting physics learning by making abstract concepts interactive, and generative AI now lowers the barrier to creating customized, embodied, interactive simulations. The work connects to Physics Education, Generative AI, and Simulation and illustrates the growing role of Prompt Engineering in educational content creation.

Lowering the Barrier to Embodied Simulation

Computer simulations have a long and well-documented record of supporting physics learning, making abstract physical concepts interactive and inviting students to play with parameters. This paper shows how a structured natural-language prompt can generate a browser-based, hand-controlled AR physics simulation.

The AR Simulation

Spread your thumb and index finger in the air, and a virtual lamp in the room changes color — a browser-based, hand-controlled augmented-reality physics simulation generated entirely from a structured natural-language prompt.

Use in Instruction

The paper describes the simulation's use in an introductory physics class, illustrating how generative AI and prompt engineering can produce customized, embodied, interactive physics learning tools without programming expertise.

Connected Concepts

Connected Articles

Citation

Levy, O., Glazer, J., Finkelstein, N. D., & Ben-Zion, Y. (2026). From prompt to embodied simulation: Using generative AI to create AR physics learning tools. arXiv:2607.24709.