Prework assignments have become essential for faculty using active learning to prepare students before sessions. However, medical students often struggle to complete prework. Non-completion has been linked to difficulty understanding learning goals and decreased session attendance. Previous work showed that the format and ease of use of prework activities greatly influence students’ motivation. Given the limited literature on audio prework, we created AI-generated podcasts (AI-GP) to improve 1st-year students’ preparation for biochemistry didactic sessions. Podcast content was created using NotebookLM, a generative AI tool. NotebookLM enabled the production of accurate, accessible podcasts from faculty-selected sources such as textbook readings and PPT slides. To prepare for 23 in-person sessions, students could listen to the podcasts, complete readings, or both for each session. A 12-question survey shared with students assessed completion rates and perceptions of AI-GP as prework material. On a 1-5 Likert scale, 72 students reported that the podcasts helped connect prework to sessions (3.7), taught foundational metabolic concepts (3.6), and that they trusted the content's accuracy (3.5). However, student engagement with prework formats was modest. Most respondents (68 of 99) listened to 5 or fewer of the 23 AI-GP, while 9 listened to 16 or more. Of the 23 readings, 60 students completed 5 or fewer, and 13 completed 16 or more readings. Qualitative analysis grouped students with low podcast completion into 3 categories: those wanting more in-depth podcasts, those with established study habits, and those opposed to AI use in education. Although students who engaged with the AI-GP found it valuable as prework, many did not develop a habit of listening and showed a slight preference for prework readings. Future work will refine the podcast structure and clearly define its role in prework to increase completion and student preparedness for active learning sessions.