The Future Is Humanist | Slow Syllabus

Humanities higher education ran on a simple loop for decades. Assign a difficult text. Expect students to struggle through. Demand an essay to prove they did the work. Repeat.

ChatGPT broke the loop in about six weeks in the fall of 2022. Suddenly anyone could generate a clean, competent analysis of Kafka without reading Kafka. Competent summaries of Kant when you can’t pronounce Kant. The prose came back polished, and it came back at scale.

The response has been predictable: universal panic at the universities. Blue book sales surged 30 percent at Texas A&M, nearly 50 percent at the University of Florida, and 80 percent at UC Berkeley. Some institutions seek detection software. Others double down on honor codes and anti-AI pledges. The impulse is to dig a moat around the classroom and throw the phones over to the other side.

But here is the hard truth: the old assignments were already failing before AI arrived. They rewarded summary over insight, compliance over conviction. Most students wrote for the assignment, not to think. The transaction that was once possible with SparkNotes and Wikipedia—or a generation before with MasterPlots—just got faster and more obvious.

The more revealing fact is this: many of our core assignments were already training students to act like chatbots. What is this reading about? Output a paragraph. What is this passage of Woolf saying? Output the correct answer. Identify the theme. Produce a clean synthesis. The highest reward went to frictionless paraphrase, familiar critical gestures, and tonal compliance. When a machine turned out to do that work better, faster, and without fatigue, it didn’t cheat the system. It completed it.

That completion should not have surprised us. Large language models are designed to generate the most statistically plausible completion of whatever has been started. The architecture optimizes for fluent pattern-matching, not accuracy, not meaning, not truth. Johan Fredrikzon, writing in Critical AI in 2025, describes these systems as “epistemologically indifferent”: they traffic neither in facts nor in fiction, only in what sounds convincingly like either. When the machine and the assignment are both optimized for the same output—a confident, polished surface with no necessary relationship to understanding—it is not the machine that has cheated. It is the assignment that has been exposed.

When I declared an English major in the mid-1990s, I was riding the discipline’s second great wave. English was near its post-Vietnam peak, conferring more than 56,000 degrees a year. To my parents, it looked like a one-way ticket to a cardboard box. To me, it felt like a ticket to infinite worlds.

The cardboard-box warning eventually won, for the statistical person, if not for me.

English majors fell from 7.6 percent of all bachelor’s degrees in 1971 to just 2.8 percent in 2021. Between 2012 and 2022, while the economy recovered, the humanities didn’t: English and history degrees fell by roughly a third. Even at Harvard, despite immense prestige and generous aid, English majors have dropped by roughly 75 percent as students chase paths perceived as more directly impactful. Students were not merely fleeing toward STEM. They were voting with their feet against a discipline that no longer made a compelling claim on their lives.

This was not simply a market correction. It reflected an internal failure of what you might call the building instinct. Over time, the humanities trained students to approach texts primarily as problems to be diagnosed rather than worlds to be entered. The hermeneutics of suspicion—the dominant mode of professional reading since at least the 1970s—became an end rather than a tool. Rita Felski has documented this tendency in clinical detail: critique became the default setting of literary study, the only move that could be taken seriously, until the discipline could no longer articulate what drew people to literature in the first place. We taught readers to knock things down more fluently than to build meaning out of them.

Meanwhile, the PhD prestige machine kept spinning. We produced a generation of brilliant people trained to work alone for six to eight years on projects no one asked for, writing for an audience of twelve. Tenure-track English positions have declined by roughly 55 percent since 2008. Fewer majors lead to fewer tenure lines, which means fewer people advising students to major, which leads to still fewer majors. Students are not irrational to avoid a discipline that cannot employ its own apprentices.

The irony is economic as well as intellectual. The dogma that STEM is the only path to a paycheck is overstated. Long-term outcomes for humanities graduates are often competitive, especially in fields that value judgment, communication, and adaptability. Employers report wanting the very skills the humanities train: ethical reasoning, synthesis, empathy. We simply stopped telling students this was what the humanities were for.

Three years after ChatGPT, the dust is settling, and what it reveals is a discipline that had been hollowing out long before the bots arrived.

The old humanities loop—read, struggle, write—depended on two contracts. A classroom contract of open inquiry, in which provisional speech and intellectual risk were welcome. And a private contract of diligent study, in which the student did the slow, uncomfortable work of reading something difficult and coming out changed on the other side.

Both contracts broke before AI. The phone came first: an external lobby for attention that made the slow struggle of difficult reading feel gratuitous. Then digital permanence: a seminar comment can now become a permanent artifact, a misstep can metastasize into a reputation, and risk becomes expensive. The result is widespread self-censorship. Students hesitate. Faculty avoid provocation. The moving encounter dies.

What remains, in too many places, is what has been described as soul-destroyed humanities: English quietly rebranded as communications, a sterile information-transfer function that a computer can, and soon will, do better.

If you can do it from behind a screen, a computer will soon do it better. If the humanities reduce to information transfer—the routine identification of familiar tropes, the production of competent summaries, the performance of critical fluency—then Wikipedia and AI have already won. The question is whether the humanities are reducible to those operations or whether they name something else entirely.

The point was never the answers. It was always the journey.

Confusing answers for meaning is the mistake that sits at the center of both the crisis in the humanities and the panic over artificial intelligence. When someone asks an AI “What is the meaning of life?” the system responds fluently. It produces language that sounds like wisdom. But it does not produce meaning on a boring Tuesday afternoon, or during grief, or in a moment of moral pressure. Meaning does not arrive as information. It emerges through repetition, commitment, consequence, and revision—over time, and in relationship with others.

Artificial intelligence drives the value of quick answers toward zero. It excels at outputs. What remains scarce, and therefore valuable, are judgment, taste, ethical reasoning, lived memory, moral accountability, and the capacity to live inside unresolved questions. Those are not defects in a humanities education. They are its core competencies.

In this light, the humanities need not be defensive. They must be foundational. When the value proposition shifts from content to meaning—from knowledge accumulation to agility of mind and strength of character—the humanities return to the center of education. In a fact-saturated world, the most important thing education can do is train people to undertake the lifelong work of becoming.

This requires thinking at two scales simultaneously. On the large scale, entire categories of cognitive labor will disappear, and new categories will emerge in their place, many not yet named. Professional pathways will be less stable, less legible, and less predictable than those higher education was designed to serve. What this world will require is not just technical flexibility but civic capacity: the ability to identify real problems, act in public, build communities, and adapt without losing one’s bearings.

On the small scale, daily life will matter more, not less. As work ceases to provide meaning by default, people will need to generate it elsewhere: in art, community, creativity, contact with the natural world, and the rhythms of everyday life. The humanities become life-centric. They train people to orient themselves when external scripts collapse.

There is an ethical implication here that is easy to overlook. Even when productivity is automated, meaning never can be. Students still need to figure out three things: what they are passionate about, what they are good at, and what the world needs. That process is deeply individual, but it cannot happen in isolation. It requires a shared community in which aspirations are articulated, tested, revised, and made intelligible to others. It requires good teachers.

This is why presence matters so much. When answers are abundant, presence becomes the scarce good. Seminar learning depends on things machines cannot replicate: embodied accountability, unrehearsed speech, risk followed by repair, trust built over time. These are not nostalgic preferences. They are design requirements for character formation.

The deeper shift required is philosophical. Critique has value, but it cannot be the destination. Suspicion is a tool, not a telos. Meaning is not something we uncover by dismantling texts until nothing remains. Meaning is composed, practiced, and sustained. Narrative is not merely an object of analysis. It is the technology of consciousness—the means by which human beings orient themselves in time, responsibility, and relationships. Education fails when it trains students only to diagnose stories rather than to live inside them and shape them wisely.

This is where the builder replaces the critic. You cannot live in a deconstructed house. You cannot be sustained if you never grow things. We have to plant something.

Practically, this means pedagogy that forces thinking into the open—and that refuses to ask questions a machine can answer for you.

In my own courses, I have stopped assigning the diagnostic essay. No more “What is this text about?” No more “Identify the theme.” Those were the assignments that were already training students to be chatbots, and AI has made their bankruptcy obvious. In their place, I assign what therapists call self-authoring: a project in which students use the semester’s texts as instruments to map the next three to five years of their lives. They identify goals across multiple domains—professional, personal, physical, spiritual, artistic. For each goal they explore what drives them toward it, what obstacles they face, how they will track progress, and what they are willing to risk. The texts we read become lenses for self-examination, not objects to be summarized.

AI cannot do this assignment for you. Not because of any surveillance mechanism, but because the assignment requires you to be a person with a past and a future, making commitments in front of other people who are doing the same thing. The seminar becomes a space of mutual accountability. Students read each other’s work, challenge each other’s evasions, and witness each other’s ambitions. The product matters less than the process. The process is the education.

[AUTHOR NOTE: Justin: consider adding one more concrete example here—perhaps a seminar exercise, something brief, that shows the in-class method as distinct from the assignment. The Neruda exercise or the vulnerability modeling would work.]

I tell my students that most of the writing they will do in college will not matter in five years. I ask them to make this project the exception. That is not a pedagogical gimmick. It is a reorientation of the entire enterprise: from proving you did the reading to using the reading to build a life.

An orchard is not a metaphor for critique. It is a metaphor for faith in the future. Cultivation presumes tomorrow. It requires patience, repetition, care, and acceptance of uncertainty. You cannot optimize an orchard. You tend it.

That tending requires a new classroom contract. Attention and risk become the price of admission. Mistakes become occasions for repair, not expulsion. These norms do not prevent cheating. They prevent shortcutting the work of becoming.

As work stops supplying meaning by default, education must teach people how to generate it—daily, locally, imperfectly, together. The humanities are the institutions explicitly responsible for that work: judgment, relationships, and meaning. They are not obsolete. They are foundational.

The choice facing higher education is not blue books versus chatbots. It is whether to cultivate an orchard or conserve the ruins. The question is not why anyone would major in the humanities. The question is how to live without them.

Justin Neuman is a Professor of Literary Studies at Eugene Lang College, The New School. He is the author of Fiction Beyond Secularism (2014) and co-author of Modernism and Its Environments (2020). He has completed over fifty trail ultramarathons and still can’t outrun his inbox.