Does Education Really Require the F-Word?
In this essay, Jon Ippolito discusses the meaning of "friction" in the context of higher education, via an exploration of what friction entails, what varieties are beneficial for students, and what aren't. In doing so, he creates the starting point for conversation in "Friction and Education: The Discussion".
Introduction
Teachers everywhere are decrying the surge in reports of students using AI to complete assignments, and they’re using a distinctive physical metaphor to explain what this leaves out of the learning process. Students are “avoiding the friction necessary for learning” by “turning to the path of least resistance”; they are “offloading the cognitive burden” onto chatbots instead of “developing grit.”
Friction has become the go-to rhetorical trope for desirable obstacles and productive struggle that help students grow intellectually and emotionally. In his rundown of arguments for preserving friction in writing (and learning in general), educator Leon Furze cites calls of alarm from a raft of educators I deeply respect (Furze 2024). One of them, Harvard Writing Center director Jane Rosenzweig, goes so far as to claim that sometimes, “friction is the point.” (Rosenzweig 2024)
I can’t blame educators for settling on the F-word, because its inverse is a staple of tech bro hyperbole. “Frictionless” has become a buzzword that sells lubricated solutions while obscuring any of their costly externalities1. The marketing of generative AI is the culmination of this rhetorical ploy, as Kyla Scanlon explains:
“AI is the newest terrain in a decades-long race to eliminate all forms of cognitive resistance… This is what a frictionless world looks like. Everything accelerates, until you forget what it means to try. Apps load faster. Papers write themselves. Job interviews are browser tricks.” (Scanlon 2025)
Before we just retire an overworked figure of speech in favor of its antithesis, however, let’s be sure we know what the word friction means. As we’ll see, not all friction builds traction; some just burns rubber.
What do we mean by friction?
Unless we’re talking about pencil lead on paper or fingernails on a chalkboard, friction in education is a metaphor based on a physical phenomenon. The laws of kinetic friction were first codified by none other than Leonardo da Vinci, who observed that the resistance to motion increases proportionally to the applied load.
It’s important to distinguish friction from the physical concept of work. When a bartender slides a pint of beer across the counter, some of the ooomph she applies with her palm does useful work getting the customer his Guiness, but another percentage of her force gets dissipated as heat thanks to the counter’s friction. So although they typically coincide in practice, friction is not the same thing as work—or for that matter, pressure, effort, or stress.
Advocates of friction in education are quick to point out that physical friction can be valuable in the right scenario; without friction, a wheel couldn’t rotate without slipping, so we wouldn’t have cars. The fact that the rough pavement drags against the tires actually helps the engine move your Toyota out of the garage. But when friction in one part of the system enables work for another, we should be asking what work is being done for whom. It’s nice that your Camry gets to speed down the street, but your tires eventually wear out from the chafing.
By analogy, students are likely to chafe at friction that is useful to teachers but doesn’t seem empowering or transformative to them. Our students aren’t monks devoting themselves to a lifetime of asceticism to achieve enlightenment; they’re unlikely to share the fetish some instructors seem to have for the path of most resistance. Many examples of friction in the classroom that seem natural to faculty are in fact unexamined traditions or accidents of history. When we notice such moments, we should be asking ourselves if they are genuinely serving student learning, or are serving something—or someone—else.
Let’s focus on educating writers. What kinds of friction are useful for them?
The friction our students probably don’t need
Immobility
Given a welter of distractions from TikTok to tumbling practice, some instructors think the best thing for today’s teens is to be forced to sit down and focus on one task for a 45- or 90-minute class. The compulsion to remain sedentary at your desk is analogous to the friction that keeps glasses from sliding off your nose—but of much more questionable utility2.
Of course, keeping students “on task” helps teachers maintain order in the classroom and helps prepare students to be industrious workers for a Taylorist economy. But I think the value to students is less than the value to their teachers or their future bosses. Friction is not useful when it is busywork that just occupies student time; it’s no wonder students pinned to a desk by drudgery get bored and want to check their phones (I would).
Attrition
Friction is also not helpful when it is a means of exerting power over others or creating social divisions. An intro course designed to “wash out” those with insufficient commitment is doing work for the administrators, not for the students. This also goes for assignments meant to weed out the underqualified, leaving only the learners with enough grit to survive the gauntlet. Your motivation for assigning a 25-page paper during the first week of class may be to show students you mean business, but they will perceive that task as capricious and cruel rather than character-building3.
Single-use learning curves
Every new task has a learning curve, but not every climb results in useful work. The toil of learning a new tool shouldn’t be a value in itself, with a possible exception for tools that will help you outlive the apocalypse by tying your shoes or scything a meadow. However, the process can be indirectly valuable if proficiency in the tool unlocks new capabilities, or what you learn can be applied to learning other tools.
Unfortunately, tool-specific friction often serves teachers who find it more convenient to use a tool in classes that has no value outside school. (That’s why I teach with Slack and Github instead of Blackboard or Brightspace.)
In the context of writing, we should ask ourselves whether tolerating cramped fingers or memorizing where the keys are on a keyboard (or even choosing between GPT-4o or o3) are truly generalizable skills. If not, that friction can be perceived rightly as wasteful, even if we had to endure it as beginning writers. Each time you ask students to learn a new tool with no generalizable benefit, it’s like asking Sisyphus to roll a new ball up the same hill.
Language mechanics
Just as Photoshop or Excel experts can be proud of their craft, so teachers adept in writing may see their self-respect bound up with a knack for choosing the right synonym or finding the perfect example for an argument. I have to admit that part of my own sense of self is coupled to how I write.
Yet if prose is a tool, could my fixation on knowing the difference between, say, a direct and indirect object be a sign that I am too wedded to this tool as evidence of my own self-worth? If so, am I more likely to clutch desperately onto this tool despite evidence that its relevance may be waning?
As I’ve argued elsewhere, the rise of large language models calls into question whether young writers still need to learn spelling, grammar, and paragraph organization in order to think. Now that non-native speakers and neurodivergent authors have a means of circumventing the mental cost of discerning comma splices or telling “it’s” from “its,” could it be that the friction required to learn the mechanics of English has become more hindrance than help?
The friction our students may need
Decision-making
Concluding paragraphs are often the hardest part of an essay for young writers because they typically require taking a stand (assuming they don’t demur to the dreaded “Both sides have a point”). Even though ChatGPT can generate conclusions with ease, the stress of making a judgment is one kind of friction I would like to preserve: it’s a generalizable task that will be required in countless high-stakes contexts later in life.
Fortunately, scaffolding the use of AI can salvage this valuable friction. Have your students ask a chatbot to generate multiple possible conclusions, then ask them to evaluate all of them and justify their choice in their own words. Let ChatGPT lay out the cases for slavery, economic differences, or state versus federal authority as the Civil War’s root cause, but make students decide between them.
Cognitive dissonance
The “choose your conclusion” exercise will have another benefit as well, namely requiring students to consider multiple viewpoints, especially those that may not come naturally. I’ve personally used chatbots to confront and challenge my theses with counterexamples or test my assumptions with different historical and philosophical perspectives, with prompts like “Offer alternative viewpoints or facts that may or may not support the argument but are missing in this draft.” Developing empathy for other people and opinions is well worth the added strain in my book.
That being said, it’s important that this critique happen at the appropriate juncture in the writing process. In “Teaching Two Kinds of Thinking by Teaching Writing,” Peter Elbow distinguishes between creative and critical stages of writing. Cross-examination can help with critical revision and reflection but can impede creative brainstorming, especially for students who already lack confidence in expressing their ideas. Sure, health experts recommend challenging yourself to boost neuroplasticity (Fisher 2025). But add too much grit in the gears and your students are as likely to give up as to rise to the occasion.
Changing context
Another example of friction I would preserve is requiring students to demonstrate the same skill with different exercises or applications. The value of this “transfer friction” has been documented in education. In one study, kids repeatedly tossed beanbags at floor targets despite a visual obstacle; for some the target was always three feet away, while others practiced at varying distances. To the researchers’ surprise, when later tested at the three-foot distance, the variable-practice children outperformed kids who had only practiced at that fixed range4. Training in different environments produces learners that are not only more adaptable, but often better even when a single task is to be optimized.
Social interaction
Most of the critiques of “frictionless AI” that I read describe the benefits of individual writers wrestling with their own ideas and words; I’ve interrogated elsewhere the presumptive benefit of writing as developing one’s own authentic voice. Now But few birds sing to themselves; as much as writing teachers often stress the importance of audience, that figure is usually imaginary (“Write an essay to persuade someone who disagrees with your position”) or at best real but asynchronous (“Write for your instructor, who will get comments back to you in three weeks”).
Actual dialogue—sharing thoughts with peers orally or via bursts of digital repartee—can be motivating for students as well as add some productive friction. Given a relatively safe space for conversation5, interlocutors with different opinions may challenge their peers to support or revise their ideas. And unlike the generic sycophancy of chatbots or neutrality of a disembodied red pen, partners in an in-person conversation have personalities that can excite, confuse, or even “rub you the wrong way.” While there are precious few examples from postgraduate life that require 5-page term papers, the ability to carry on a productive conversation with neighbors or coworkers you disagree with is an absolutely essential skill in today’s workplaces and social spaces, whether online or off.
It’s important to note that a thoughtless use of AI will probably shortcircuit this kind of friction. In his book Cointelligence, Ethan Mollick recounts asking why students don’t raise their hands anymore, only to be told there’s not much point when they can just consult ChatGPT at home. The incident is mentioned in a recent episode of the Teaching in Higher Ed podcast, where Carter Moulton contrasts the commonplace call for a human-in-the-loop solution to AI’s dangers with a more sociable approach he calls humans-in-the-loop: “All of a sudden we have more opinions, we have more disagreements; more friction that can help the learning process, but also bring folks together.” (Moulton 2025)
Conclusion
I’ve argued above that some forms of friction in education—like immobility, attrition, single-use learning curves, and even language mechanics—can be counterproductive. I’ve also argued that other kinds can galvanize the learning process, such as requiring decision-making, cognitive dissonance, changing context, and social interaction. Now that large language models remove many of the obstacles in the way of fluent prose, I find myself increasingly thinking about these nested sets:
-
All frictions produced by mastering prose.
-
All frictions in set 1 that are productive.
-
All frictions in set 2 that can only be experienced by constructing grammatical sentences and fluent paragraphs.
I believe we writing instructors need to think long and hard about what’s actually in set #3, and whether those gains are worth the counterproductive frictions that come along for the ride.
Friction can enhance learning, but only when it’s applied with a clear purpose. If I’m going to ask writers to struggle, I owe them an honest explanation of how the added wear will benefit them. Otherwise we may be forcing students to bear the brunt so that their teachers (or future bosses) won’t have to—and there’s the rub.
References
Bjork, Elizabeth Ligon, and Robert A. Bjork. “Making Things Hard on Yourself, but in a Good Way: Creating Desirable Difficulties to Enhance Learning.” Psychology and the Real World: Essays Illustrating Fundamental Contributions to Society, Worth Publishers, 2011, pp. 56—64.
Colombia Doctors. “Take Your First Step Towards Better Health with Walking.” ColumbiaDoctors, 1 Feb. 2024, https://www.columbiadoctors.org/news/take-your-first-step-towards-better-health-walking.
Duran, Andrea T., et al. “Breaking Up Prolonged Sitting to Improve Cardiometabolic Risk: Dose—Response Analysis of a Randomized Crossover Trial.” Medicine & Science in Sports & Exercise, vol. 55, no. 5, May 2023, p. 847. journals.lww.com, https://doi.org/10.1249/MSS.0000000000003109.
Ippolito, Jon. “Writing As Thinking—By Proxy.” electronic book review, 18 Feb. 2026. https://www.electronicbookreview.com/publications/writing-as-thinking-by-proxy/
Fisher, Jennifer. “Tips to Leverage Neuroplasticity to Maintain Cognitive Fitness as You Age.” Harvard Health, 2 Apr. 2025, https://www.health.harvard.edu/mind-and-mood/tips-to-leverage-neuroplasticity-to-maintain-cognitive-fitness-as-you-age.
Furze, Leon. “Why Writing Needs Good Friction.” Leon Furze, 3 Jul. 2024, https://leonfurze.com/2024/07/04/why-writing-needs-good-friction/.
Moulton, Carter. “Analog Inspiration: Human Centered AI in the Classroom with Carter Moulton┬╗ Teaching in Higher Ed, Episode 593, October 23, 2025, https://teachinginhighered.com/podcast/analog-inspiration-human-centered-ai-in-the-classroom-with-carter-moulton
Rosenzweig, Jane. “When the Friction Is the Point.” Substack newsletter. Writing Hacks, 29 May 2024, https://writinghacks.substack.com/p/when-the-friction-is-the-point.
Ris, Ethan W.. Grit: A Short History of a Useful Concept | MABEL | Western Washington University, 1 January 2015 https://mabel.wwu.edu/do/94a7219b-a1cc-435d-a3f2-be68710d4a4a.
Scanlon, Kyla. “The Most Valuable Commodity in the World Is Friction.” Substack newsletter. Kyla’s Newsletter, 8 May 2025, https://kyla.substack.com/p/the-most-valuable-commodity-in-the.
Footnotes
-
The difference between work and friction discussed later in this essay—and how so-called frictionless systems often externalize their wastefulness in hidden contexts—also applies to the hardware that runs generative AI. With an object in motion, friction converts some of its kinetic energy into heat; with current zipping through a circuit, resistance dissipates some of its electrical energy as heat. This is why the GPUs processing your AI queries in a data center require prodigious amounts of water to cool off. For more on the water usage of data centers, see the author’s What Uses More app, which draws on publicly available estimates and lets users tweak parameters such as prompt complexity or the energy source and climate of local data centers. https://What-Uses-More.com. ↩
-
Getting your butt out of your chair even for a matter of seconds will not just keep you awake, but can have measurable health improvements. A Columbia University study detected better blood markers when adults stood up for just 1-5 minutes every 30 minutes. (Duran et al. 2023) Such “movement breaks” also dramatically affected how the participants responded to large meals, reducing blood sugar spikes after eating by 58% and reducing fatigue by 25% compared to sitting all day. (ColombiaDoctors 2024) ↩
-
When I surveyed successful alumni of the University of Maine’s New Media program in 2021, they ranked grit lowest out of ten soft skills required for success. Historians have also noted that use of the term has an ulterior motive in rationalizing class privilege. “Grit is an eminently useful concept, but not because it can help the prospects of disadvantaged students. Instead, it helps middle and upper-class adults explain and counteract the shortcomings of their own children, and it also helps them put off the sacrifices that could break down the American caste system.” (Ris 2015) ↩
-
“Eight-year-olds and 12-year-olds practiced throwing beanbags at a target on the floor with their vision occluded at the time of each throw. For each age group, half of the children did all their practicing throwing to a target at a fixed distance (for example, 3 feet for the 8-year-olds), while the other half threw to targets that were closer or farther away. After the learning sessions and a delay, all children were tested at the distance used in the fixed-practice condition for their age group (Kerr & Booth, 1978). Common sense would suggest that the children who practiced at the tested distance would perform better than those who had never practiced at that distance, but the opposite was true for both age groups.” (Bjork and Bjork 2011) ↩
-
Anecdotally, my children and their friends in US public schools have reported that the intellectual environments they encountered in high school and college in the 2020s were often “too safe,” by which they meant teachers steered conversations away from controversy when discussing volatile subjects. To be fair, this tradeoff between maintaining classroom civility and free speech is a difficult balance to achieve, whether in online or physical spaces. ↩
Cite this essay
Ippolito, Jon. "Does Education Really Require the F-Word?" Electronic Book Review, 18 February 2026, https://electronicbookreview.com/publications/does-education-really-require-the-f-word/