Artificial intelligence is increasingly integrated into daily life, and modern educated individuals should have the ability to use AI tools correctly to improve work, study, and life efficiency. In this context, artificial intelligence literacy has been proposed. Due to the lack of consensus on the constructs of artificial intelligence literacy, this study used the scoping review to summarize the AI literacy constructs, including recognize AI, know AI, AI ethics, AI empowerment, AI self-competence and apply AI. In order to further explore the relationship between these six constructs, this study distributed an artificial intelligence literacy questionnaire to 276 non-expert university students (referring to students who have not received formal artificial intelligence education) and used structural equation modeling to verify the hypothesis. Research has found that recognize AI, know AI, AI ethics, AI empowerment, and AI self-competence all have significant positive predictive effects on apply AI. Know AI also has a significant positive predictive effect on AI ethics, AI empowerment, and AI self-competence. AI ethics, AI empowerment, and AI self-competence play a mediating role in the relationship between know AI and apply AI. The findings further improve the constructs exploration of artificial intelligence literacy in current research and provide some inspiration for teaching practice.