The development of artificial intelligence (AI) technology has made the development of AI literacy a pressing need. However, the current AI literacy assessment methods for college students are mainly traditional methods that consider reliability and hysteresis, such as questionnaires and interviews. Few studies have made use of topic modeling techniques. This study explored the AI literacy that different students demonstrated in their AI course assignments and assessed the relationships between different AI literacy topic words. The results highlight that “online,” “learning,” “students,” and “education” are the most often used topic words in relation to AI literacy. This suggests a strong connection and possible overlap in discussions about AI literacy. In the future, these topic words could be regarded as a reference to improve AI literacy for students of different disciplines and age levels.