You've heard of Google NotebookLM and its "audio overview" feature, right? This is the AI-powered tool that can create a podcast-style audio summary of one or more files submitted to NotebookLM. Within a minute or two, you can hear two chatty podcast hosts talk to each other about whatever reading you've given them. I've heard faculty worry that students might start listening to NotebookLM's summaries of course texts instead of doing the hard work of actually reading the course texts. How will students develop much-needed reading skills if they outsource the reading to a couple of chatty robot podcast hosts?

I haven't been too worried about this, likely because the first time I heard a faculty member share their actual experience with NotebookLM it was an education professor whose independent study student had found those audio overviews a helpful scaffolding for reading the challenging journal articles the professor assigned her. The student would first listen to the "podcast" version of the article to get oriented to the main ideas and arguments and then dive into the article itself to make sense of the details. Just this week, community college instructor Michelle Kassorla posted about a student in her course using NotebookLM in the same way, as a pre-reading activity.

Not all students will use NotebookLM so thoughtfully, of course. But might we teach our students to use AI summarizers as a kind of scaffolding for reading challenging texts? Maybe students should try reading the original text without AI assistance first and then consult an AI overview to see if it aligns with their initial understanding? Or maybe any AI involvement is likely to lead to a deskilling so that we should encourage students to skip the AI and lean into the difficulty of reading?

I was thinking about this last week while talking with Jim Lang for an upcoming episode of Designed for Learning, the podcast he hosts for the University of Notre Dame. Jim asked me about brainstorming and not reading, but his question was essentially the same: If students bring AI into the process, do they risk some kind of deskilling? Most of us have experienced the way that reliance on GPS for navigating makes it harder to learn the roads and streets of a city. Might reliance on AI summaries similarly inhibit development of reading skills?

Sure, maybe. But don't we already give students a bit of scaffolding for challenging course readings? There's that mini-lecture the class session prior where we provide some history and context for the reading. There's the reading response essay we ask students to write as a reflection on the reading. And there's the class discussion of the reading, which we hope deepens student understanding. It's not like we ask students to go off and do the reading completely on their own. We design a set of learning experiences that we hope will assist students in the reading process.

What if it's the class discussion that's inhibiting the development of student reading skills? Or those reading response prompts we provide? Or the pre-reading lecture? If we take those "crutches" away, do students still leave our courses knowing how to read the challenging texts in our disciplines?

Y'all know I'm a fan of social annotation. I love asking students to "do the reading" together in some digital space where they can highlight and comment on various passages and engage in discussions with each other that are anchored in the text itself. I find that this is a more authentic way to engage with a reading than asking students to comment on discussion board threads. But am I doing my students a disservice if I provide this kind of scaffolding? Will they be able to read hard texts without a class full of peers annotating the text together?

I don't think social annotation is a crutch. I don't think class discussions inhibit students' reading skills. And I don't think AI summaries are a problem, either, if used thoughtfully. When we direct students to these activities, we're actually directing students to strategies they can adopt to improve their reading skills. We're teaching them practices they can use after the course is over when they need to read and understand challenging material.

Consider the ways that you tackle a journal article or book chapter or other reading in your work. What annotation techniques do you favor? Who are the colleagues you talk to about the reading? How do you seek out background information you might need? What pre-reading steps do you take? I imagine you have a well developed set of techniques for doing this work. I do, and now occasionally that set of techniques includes an AI summary.

Perhaps we should think about the various reading activities we give students not so much as scaffolding for their current reading experience but as practice for future reading experiences. By asking students to listen to a background lecture or annotate with colleagues or write a reflection piece or have an AI chatbot ask them questions about a reading, we're showing students a variety of strategies they can adopt in the future for the reading they need or want to do.

I don't always use the same reading strategies for everything I need to read. The nature of the text, my goals for the task, the contexts in which I'm reading--all of that helps me decide which strategies to employ. I'd like my students to have a healthy toolbox of reading strategies and to know which ones work for them for different reading purposes. That's a good reason to structure different activities into a reading-focused course, showing students the range of tools they might put in their toolbox.

Getting reading assistance from AI is strange and new, and we all know AI gets stuff wrong sometimes. So we ask students to document and reflect on their use of AI in a reading activity. That reflection is a core element of a "green light" AI assignment. The hope is that the reflection will help students better understand the strengths and weaknesses of the AI assistance, which in turn will help them make better choices about when and how to use AI in the future.

What if we took the same approach to these other reading activities? Ask students, "List the people you discussed this week's reading with and note what you gained from these conversations." Or ask, "If the goal was a deeper understanding of this week's reading, was the time you spent annotating and reading your peers' annotations a good investment?" Or ask, "In what ways did the pre-reading lecture help your close reading of this week's text?" (Or maybe don't ask that.)

When we try to identify the one best way for students to go about reading in our fields, it's doubtful that a podcast-style AI-generated audio summary is going to make the cut. But if that particular form of AI assistance is but one of many strategies that are potentially useful depending on the reading and the student and the student's purpose in reading, then our job becomes helping students learn when and how to deploy these various strategies. That's not an easy job, but it gives us a way to think about new technologies like generative AI and their potential for helping students learn to read deeply.

Thanks for reading!

If you found this newsletter useful, please forward it to a colleague who might like it! That's one of the best ways you can support the work I'm doing here at Intentional Teaching.

Or consider subscribing to the Intentional Teaching podcast. For just $3 US per month, you can help defray production costs for the podcast and newsletter and you get access to subscriber-only podcast bonus episodes.