NotebookLM Infographics: Ultraprocessed, Biased Slop

Ok, I’ll bite: First take with the new NotebookLM infographic feature: Surprising? Sure. Slop? Yes. Biased and mostly useless? Definitely.

First, context: I have a Notebook called “Remi’s Annotation Research” that contains 46 sources. It’s all my original scholarship, articles and chapters and posts and notes, from nearly a decade (but not, intentionally, my latest book). Whenever NotebookLM introduces new features, I use this collection for initial experimentation.

So, my prompt to generate this infographic:

“Drawing upon my scholarship, create an infographic about the educational benefits of annotation, highlighting how this literacy practice affords social, collaborative, and critical learning. Create the infographic in the visual style of a medieval manuscript with elaborate rubrication, ideas emphasized with marginalia and interlinear notes, and include some fantastical animals that are discussing main ideas.”

“More than Margins” indeed: Griffins! Dragons! Unicorns! Snakes!

My surprise lasted a minute.

But look carefully beneath the synthetic shine and note how racial and gender biases are visualized in this image, how dominant social and epistemological hierarchies are reified, all despite the fact that my scholarship about annotation demonstrates the opposite and more critical affordances of this practice.

Here are four examples:

1. Yes, annotation “Makes Learning a Shared Dialogue.” It’s notable, however, that the dialogue occurs between an older white sage and a younger white pupil. Why?
2. Yes, annotation is a means by which a reader-as-annotator “Joins a Centuries-Old ‘Great Conversation.’” But conversation, apparently, only occurs among white male learners, from a scribe through to the kid holding a tablet. Seems my writing about annotation and/as citational justice wasn’t noteworthy enough for this algorithm to visualize. 
3. The idea “Critiques Power and Rewrites Narratives” is illustrated by an old white king whose quill wrote a red “X.” I write extensively about the power of “X” as an annotated symbol in my book “Re/Marks on Power;” briefly, that annotation was used by Harriet Tubman in her advocacy, and more recently by educators like Debbie Reese to contest inaccurate representations of Indigenous peoples. How disappointing that this image, generated in reference to my scholarship, associates such a powerful mark of resistance with such a conventional historical authority.
4. It’s also troubling that the only figure with a speech bubble of text is both Black and speaking gibberish: “Empowers shemd projects their annotated thoughts.” Given how few textual inaccuracies appear in the infographic, it’s problematic that this obvious error is conspicuously voiced by a young Black learner. 

But it’s just one image… I could keep prompting the algorithm, generating endless interpretations… but no, this single take is so problematic! It’s literally illustrative of coded bias. It’s an image I don’t want associated with my work. 

An infographic should simplify complex ideas and facilitate communication. Generously, this image might do that. But this image wouldn’t aid in information retention or assist with decision-making about research. I don’t want my scholarship summarized as superficial media. More importantly, I don’t think student engagement with ideas summarized as visuals like these will sustain their interest or scaffold their learning. I believe our students deserve better if we’re going to tinker with AI for their potential benefit—and that means higher-quality educational resources, more accurate summaries that lead students back to primary sources, something that will pique their intellectual curiosity to ask deeper questions.

NotebookLM does some things really well and I’ve given the tool props when it has succeeded at providing solutions to genuine educational problems (read this post). However, as visual summaries of primary sources, the infographics I’m currently seeing are AI slop akin to ultraprocessed food manufactured in a lab to maximize consumption and profit at the expense of nourishment and health. As a scholar and educator, I’ll pass on this ultraprocessed facsimile of knowledge.