Kwun Tong Bay Secondary School, Kowloon, Hong Kong SAR, China

Abstract: Since the introduction of virtual experiments, augmented reality, sensors, and generative artificial intelligence into Hong Kong junior secondary Science/Physics classrooms, an implicit regression has emerged: students can "do correctly" free fall, series circuits, and refracted light paths in simulation software, yet find it increasingly difficult to explain why a real balance scale wobbles, why a real ammeter's reading jumps, or why a real inclined-plane trolley veers off track. Writing from the front line as a Hong Kong junior secondary physics teacher, this paper proposes the "Sim2Real Dual-Track Experimental Model": the simulation environment is not used to prove formulas, but to "nurture hypotheses, nurture variable awareness, and nurture aliases of error"; the real experiment is not used to repeat the textbook, but to "hit errors, revise models, and explain why reality refuses to behave." Three lesson cases—S1 optics (refraction), S2 series/parallel circuits, and S3 kinetic energy and friction—illustrate that the highest value of digital education in physics is not to make physics easier, but to let students see earlier that "the simulated world is too clean; the real world is what demands thinking." The model aligns with the 2025 Hong Kong "Science (S1–S3) Curriculum Framework," specifically its directions of "scientific investigation, scientific modelling, digital education tools, and applications of artificial intelligence," and can be implemented in ordinary classes using only mobile-phone sensors and free simulators.

Keywords: educational digitalisation; junior secondary physics; virtual experiment; real experiment; Sim2Real; scientific modelling; Hong Kong Science Curriculum