Objective: This study aims to map students' Gen AI usage profiles during physics learning. In addition, this study will also examine the correlation between these variables and students' critical thinking skills in physics learning. Method: A non-experimental, quantitative, ex post facto design was used in this study. This research design captures phenomena that occur naturally without treatment. The research instrument used questionnaires and tests. Data analysis used descriptive statistics, the Kruskal-Wallis test, and the Spearman correlation test. Results: Students still tend to accept answers from Gen AI quickly during physics learning. Only a small portion are accustomed to doubting, checking, and evaluating Gen AI answers. There are differences in critical thinking skills across groups of students with different patterns of Gen AI use, with an ε2 of 0.410. The profile of Gen AI usage is positively correlated with critical thinking skills, with a correlation coefficient of 0.525 in the medium category. Novelty: This study contributes by revealing the profile of students' Gen AI use in physics learning and by providing empirical evidence that the quality of ethical and reflective use has a strong influence and is positively correlated with critical thinking skills.