<![CDATA[AI-supported online physics learning can expand access to explanations, but it may also increase cognitive offloading and uncritical reliance on generated answers. This study examined whether human-in-the-loop (HITL) AI-supported scientific argumentation improves critical thinking, scientific reasoning, computational thinking, and physics problem solving. A quasi-experimental pretest-posttest non-equivalent control group design involved 180 undergraduate students from Universitas Islam Negeri Raden Intan Lampung assigned to online argumentation without AI, AI-only learning, or HITL-AI argumentation. Primary empirical data were analyzed using normalized gain, one-way ANOVA, pairwise t-tests, Cohen's d, and Pearson correlations. The HITL-AI group obtained the highest posttest performance and medium normalized gains across outcomes (g = .38–.55). Group differences were significant for all N-gain variables (p < .001), with large effect sizes (η² = .54–.75). HITL-AI also showed higher argumentation quality, metacognitive monitoring, and germane load, alongside lower AI overreliance and cognitive offloading than AI-only learning. AI integration in physics learning should be structured through instructor-mediated argumentation protocols to preserve epistemic agency and strengthen higher-order thinking.]]>
Copyrights © 2026
