I uploaded notes on Aristotle’s Virtue Ethics to Google’s NotebookLM. The resulting audio is below. The notes I used are lengthy and detailed, so the AI had to focus on some areas to the exclusion of others. I tried a couple of versions of audio. Each one was slightly different. They also tended to contain an inaccuracy to two. However, this was likely in part because the notes were not optimised for NotebookLM’s use.

To any A-level students listening, remember that although listening to audio is a great way to learn, you need to do something with what you are listening to for it to stick. This audio is best used as an introduction to the topic or as part of a robust revision strategy.

I plan to produce more of these in the future, including ones that home in on specific aspects of Aristotle’s Virtue Ethics.

I think the results are impressive. Is technology like this a soon to be forgotten fad or something that will be of substantial use in the future? I think the latter, but of course, there are pitfalls.

Audio-producing AI have can make accessible summarise of information with remarkable ease. Podcasts and audiobooks are popular as they can be listened to no matter what someone is doing: on the morning commute, making dinner, completing mindless admin tasks. As a revision devise, I can see how audio can is a useful ways of providing a quick overview. If students hear something they don’t fully understand or remember then they can look into it in more detail.

But this points to the limitations of audio. We like it because it is easy, but learning shouldn’t be too easy. How often you listen to a podcast then find a couple of says later that you have forgotten the majority of what you have heard? Of course, we can mitigate these negative effects by doing something with the information we have heard, like writing it down or explaining to to someone. Learning requires action. Deep learning involves some challenge, reflection and concentration. So long as we are aware that listening to audio like the one I’ve produced is likely to produce only shallow learning if not combined with other approaches, I see no reason why such technology cannot play a useful role in a teacher’s toolkit.