AI-generated podcasts as supplementary module support: A short scholarly reflection

For my upcoming teaching in Strategic Project Management and Industrial Work Design and Organisation, I have developed a series of short audio podcasts using Google NotebookLM. A previous set for the now retired Strategic Supply Chain module remains available on the same YouTube channel. In each case, the process has been fairly simple: I upload my lecture slide decks into NotebookLM, generate an Audio Overview, and make the result available to students through the university portal alongside the rest of the module materials. The podcasts are short, usually no longer than 20 minutes per lecture, and are intended to give students another route into material that would otherwise sit mainly in slide form. Google presents Audio Overviews as generated discussions grounded in uploaded sources, which is broadly consistent with how I have used the tool here (Google, 2024, 2025).

The main value of this work lies in terms of access and flexibility rather than novelty. Students do not always return to lecture materials under ideal conditions, and they do not always find slides the easiest format through which to revisit complex ideas. CAST’s Universal Design for Learning Guidelines start from the premise that there is no single means of representation that will be optimal for every learner and that educators should provide options for how information is presented and revisited (CAST, 2024). On this basis, these podcasts can be understood as one additional representational layer around existing module content. They do not replace lectures, seminars, reading, or discussions but offer another way of encountering material that has already been designed for the module, and this is a more proportionate claim.

This claim also fits reasonably well with the existing literature on recorded and audio-based learning resources. Morris et al. (2019) describe lecture recording as a contested space between students and teachers, but their findings also show that students place clear value on recordings as learning resources. Nightingale et al. (2019) reach a more specific conclusion in relation to inclusion, arguing that supplementary lecture recording can support students with dyslexia and other specific learning difficulties and can contribute to more inclusive curricula. My NotebookLM podcasts are not lecture capture in the usual sense because they are generated summaries based on slide decks rather than recordings of live teaching. Even so, they sit within the same broad family of supplementary resources that help students flexibly revisit taught material (Morris et al., 2019; Nightingale et al., 2019).

The podcasting literature also helps locate this practice more precisely. McGarr (2009) proposed that educational podcasting in higher education tends to work in substitutional, supplementary, or creative forms. The supplementary category is the closest fit here: these podcasts are not intended to replace classroom teaching and are not designed as student-created assessment artefacts. They are brief supporting resources linked directly to existing lecture materials. A more recent scoping review by Thomas et al. (2025) shows that podcasts in higher education are now used across a wide range of teaching and learning contexts and notes their potential to give learners more choice over when, where, and how they engage with content. This is useful in this case because portability is one of the clear practical strengths of short audio summaries. Students can listen while travelling, walking, or preparing for class, and that convenience may make it easier for some of them to re-enter content they might otherwise postpone.

What makes this approach more defensible, in my view, is that it remains tightly source-bounded. I am not asking a general AI system to invent teaching content from an open prompt. I am using NotebookLM to re-present material that I have already designed in my lecture slides. That distinction is important because it maintains academic control at the level of source design. The podcasts are therefore better understood as an alternative rendering of an existing curriculum than as an independent teaching voice. Google has also cautioned that Audio Overviews are generated outputs and should not be treated as comprehensive or objective accounts, which is an important limitation to keep in view (Google, 2024, 2025). The educational role of the podcasts is therefore supportive rather than definitive.

This limitation should not be forgotten: an AI-generated podcast is not the same as teaching; it cannot respond to questions, notice confusion, adapt explanations in real time, or exercise professional judgement in the way a lecturer can during a live session. It also inherits both the strengths and weaknesses of the source material on which it is based. For this reason, I do not see these podcasts as a substitute for live teaching or direct engagement with lecture slides and other resources. I see them as one small part of a wider support structure around the module; their value lies in extending the life of existing materials and making them available in another mode, not in replacing the relational and interpretive dimensions of academic work.

Overall, this felt like a credible but modest use of generative AI in teaching. The educational claim is limited, and that is part of its strength: the podcasts are short, optional, tied directly to lecturer-produced materials, and clearly supplementary in purpose. They may help some students revisit key ideas in a form that is portable and manageable, especially when time, attention, or study conditions make slide-based reviews less appealing. For me, that is a defensible use of a tool like NotebookLM in higher education; not as a shortcut around teaching, but as a bounded way of widening access to content that remains academically designed, contextualised, and owned by the lecturer (CAST, 2024; Google, 2025; Thomas et al., 2025).

References

CAST. (2024). UDL guidelines 3.0 https://udlguidelines.cast.org

Google. (2024, September 11). NotebookLM now lets you listen to a conversation about your sources. https://blog.google/technology/ai/notebooklm-audio-overviews/

Google. (2025). Generate Audio Overview in NotebookLM. https://support.google.com/notebooklm/answer/16212820?hl=en

McGarr, O. (2009). A review of podcasting in higher education: Its influence on the traditional lecture. Australasian Journal of Educational Technology, 25(3), 309–321. https://doi.org/10.14742/ajet.1136

Morris, N. P., Swinnerton, B., & Coop, T. (2019). Lecture recordings to support learning: A contested space between students and teachers. Computers & Education, 140, Article 103604. https://doi.org/10.1016/j.compedu.2019.103604

Nightingale, K. P., Anderson, V., Onens, S., Fazil, Q., & Davies, H. (2019). Developing the inclusive curriculum: Is supplementary lecture recording an effective approach in supporting students with specific learning difficulties (SpLDs)? Computers & Education, 130, 13–25. https://doi.org/10.1016/j.compedu.2018.11.006

Thomas, M. A. M., Boivin, K., Brehm, W., Aktas, F., Karaagacli, B. C., Cho, I., & Mensah, P. K. (2026). Podcasts as pedagogy in higher education: A scoping review to map and advance the field. British Journal of Educational Technology, 57, 377–394. https://doi.org/10.1111/bjet.70021