The physicality of learning
Generative AI threatens the valuable process of searching for answers
Higher education learning used to be as much about the process as it was any answer. Learning was physical, tangible, and immersive. It required travel—to archives, libraries,1 and forgotten corners of institutional storage—and it required patience, leafing through pages, following one citation to the next, waiting for interlibrary loans to arrive. Most importantly, it cultivated a particular kind of intellectual intimacy with the material that is increasingly under threat in an era where algorithms, and now, generative AI, deliver answers on demand.
AI holds the potential to make research more efficient and learning more accessible, but the problem with efficiency is that it prioritises outcomes over process, answers over engagement. The laborious act of sifting through stacks of books, meticulously following footnotes, and struggling through faded handwriting on fragile documents was never just a means to an end; it was the learning itself. In handling physical materials, researchers absorbed the context of knowledge, the surrounding discourse, and the implicit debates that AI’s summarised responses obliterate.
The notion that research and learning should be frictionless is, at its core, a misunderstanding of how knowledge is formed. Consider the scholar working through a set of archival letters. The process of unfolding each brittle page, deciphering inconsistent handwriting, and cross-referencing dates and names is not just an inconvenience; it is an engagement with the intellectual world of the subject. These physical encounters create a depth of understanding that AI-generated summaries, no matter how sophisticated, simply cannot replicate.
There is also something to be said for serendipity. Those who have spent time in libraries know that some of the most meaningful discoveries are accidental. You search for one thing and stumble upon another—a marginal note, a forgotten reference, an unexpected connection or colourful cover that alters the course of your research. Digital tools, which give you “precisely” what you need, remove this unpredictability. They reduce learning to a linear path rather than the tangled, often frustrating, but ultimately enriching experience it should be.
What happens when we are no longer forced to engage with sources on their own terms? Research and learning have always been shaped by their physical constraints—by what was available, what could be found, and what could be read in the time one had available to them. These constraints were not limitations but structuring forces, shaping the intellectual projects of entire generations. AI, by providing instant access to summarised knowledge, removes these constraints but does so at the cost of serious engagement, of process as learning—of intimacy.
If knowledge is too easily accessed, it is not fully absorbed. A student who compiles a literature review by prompting ChatGPT for article summaries does not develop the same nuanced understanding as one who has actually read, struggled with, and contextualised each source. The former gets the information; the latter gains expertise. It is the difference between possession and comprehension.
There is a danger, too, in how AI restructures our relationship with authority. In traditional research, the process of gathering sources and verifying claims builds an inherent criticality—one must decide which sources to trust, which arguments hold weight, and where biases might be lurking. AI, by contrast, collapses all of this into an answer delivered in a neutral, confident tone, eliding the necessity of doubt. The authority of the machine is one that discourages interrogation.
Yes, yes, generative AI is powerful and can and will do much good, but as scholars, we must resist the creeping idea that research and learning should be effortless. The physicality of learning—the movements, the handling, the slow work of reading—is not just an antiquated inconvenience. It is a formative intellectual practice, one that shapes not just what we know but how we come to know it. If we lose that, we lose more than just a method—it is hard to quantify exactly what that is, but it seems like an awful lot.
For younger readers, libraries were places where subject specialists used to curate well-informed books and other knowledge resources; if you needed to develop an in-depth understanding of a particulary topic, you could go read about it at the library.