The current issue of The Atlantic proclaims the demise of reading—the act at the heart of the humanities. While the decline in reading—and the study of the humanities in general—is nothing new,[1] the Atlantic piece sparked an avalanche of responses (for a quick sampling, just look at Reddit). The piece and the reactions raise a broader set of issues: In the age of AI, what justifies an education grounded in the humanities?
For some of us, this is déjà vu. Decades ago, as the director of arts education at the National Endowment for the Arts (NEA), I was tasked with making the case for including the arts in public education. Two arguments stood out: Arts education improves math scores! The arts are a major sector of economic activity! I regularly received the latest studies to bolster both claims. If I were still making those arguments today, I could cite a long list of studies linking arts education to a variety of positive academic and social outcomes, and cite evidence that the arts generate $1.2 trillion in economic impact—over 4 percent of U.S. GDP.
Although I confess to having used both arguments in appropriate venues, they make me uneasy. First, we never justify studying trigonometry by arguing that it improves artistic creativity. Why do we ask the arts to do the reverse? More importantly, the studies cited are almost always correlational rather than causal. A sober review of the evidence by a team sympathetic to arts education found the hard evidence of impact to be minimal. The largest and most recent rigorous study showed a very small impact on writing skills and zero impact on math learning. No study has shown that replacing a math class period with one in the arts would lead to higher math scores. When it comes to the economy, the definition of arts education used in the literature is so broad as to border on meaningless. The study cited above reports that “the nation’s largest arts industry in 2023 was web publishing and streaming, which includes internet publishing and broadcasting, music and film archives, comic syndicates, and news photo distribution services.” The link to a school orchestra or painting class is not immediately evident.
I rehearse these arguments today to show why contemporary defenses of a humanities education leave me equally unpersuaded[2]. The form of the argument has shifted. In the age of AI, we are most likely to hear that the liberal arts are the best gateway to the so-called “durable skills” our new economy depends on. A recent piece in Forbes put it this way:
We need a better way of talking about the important skills that these [humanities] programs develop. They are the durable skills that employers overwhelmingly value, such as critical thinking, problem-solving, and teamwork. And they are the skills that will help people adapt and innovate as artificial intelligence reshapes work…. Metacognitive skills will be very important—flexibility, adaptability, experimentation, thinking critically, being able to challenge things.
If only it were true. The problem with these kinds of statements is that they rest on a far weaker evidentiary base than the arguments for arts education’s contribution to math learning or economic impact. Claims that a liberal arts education is especially important in the age of AI often center on the idea that it cultivates critical thinking, but the empirical basis for that claim is extremely weak. Available studies more often show correlation than causation, and I am not aware of any rigorous evidence demonstrating that humanities or social science education produces stronger critical thinking than science or math training.
The explanation for this is not complex: In almost no cases do skills developed in one academic domain transfer to others. Being excellent at assessing how Shakespeare’s King Lear self-destructs doesn’t give one a transferable skill in critical thinking that will magically manifest itself ten years later while coding for Claude.
If you ask AI whether classicists make good businessmen, it will tell you they do—partly because they are said to develop strong analytical skills. But that’s a classic (sorry) case of confusing correlation with causation: those who study classics are likely to have strong analytical skills. More to the point, there is no rigorous evidence that classicists are more likely to become terrific entrepreneurs than those who studied history, science, or math.
The second defense of a humanities education is perhaps best articulated by Martha Nussbaum. In multiple books[3], she argues that studying the humanities is crucial for developing the skills of democratic citizenship. Once again, it’s about critical thinking—with the addition of cultivating the imagination and ethical sensibilities. Once again, there is next to no evidence. As Kate Fullagar wrote last year in Inside Story, “she never deals with the unspoken question running through the book of why other disciplines don’t lead to critical thinking or imagination.” More fundamentally, Fullagar quotes Niall Lucy in his own review of Nussbaum: the humanities are “not inherently democratic, which is why fascist regimes could take root in countries (Spain, Italy, Germany) with rich literary histories.” Or as my father put it more succinctly, “Goethe’s garden is a few thousand yards from Buchenwald.”
So, what should we teach in the age of AI? Rick Hess recently offered a strong, commonsense response—just hold steady on traditional content. Hess dismisses economist Tyler Cowen’s claim that “the curriculum itself is now radically obsolete,” rejecting Cowen’s substitution of pseudo-knowledge for the real thing. We can buttress Hess’s argument—with the same note of common sense—by observing that you can’t ask AI advanced questions or work with its answers unless you know something about the field in which you are asking. One recalls William Blake’s phrase “the idiot questioner”—garbage in, trivia out.
But even fleshed out, Hess’s common sense still offers no special argument for an education in the humanities. For that, let’s pose a question: Would you give back those elements of your own education? Put another way, who among us, privileged to have been exposed to music, art, poetry, drama, philosophy, and foreign languages, would willingly undo that education? Would we abandon the education that J.S. Mill called “the strengthening, exalting, purifying, and beautifying of our common nature”? It is often said that privileging an education in the humanities is elitist—but the argument is exactly wrong: It is the exclusion of so many students from that education that is elitist and immensely patronizing.
In a culture that champions rapid consumption and disposal, an education in the humanities appeals to the lived experience of being removed from the everyday, of undertaking encounters that bring cognitive and emotional dislocation, searching self-questioning, and often joy.
If children responded only to an education that is “relevant,” Harry Potter would have remained unread. Instead, they discover what it means to lose themselves in imaginary worlds. In a time of often overwhelming social noise, knowing passages of inexhaustible poetry by heart creates inner spaces of private treasures. The encounter with art becomes not an interruption before lunch but a site of intense interrogation. When British scholar Peter de Bolla finally learns to “see” Barnett Newman’s Vir Heroicus Sublimis and discovers “a heightened sense of being well-composed,” he is speaking not of the painting but of himself.[4]
De Bolla’s experience underscores the burdens and rewards of study. Perhaps the Mona Lisa is instantly “accessible,” but many of the brightest jewels of the humanities aren’t microwave-ready. As with Bach’s Goldberg Variations, so with Dante’s Inferno: the first encounters might elicit pleasure—but unpacking the structure of the canons through repeated listening in the former or tracing the analogical echoes in a canto in the latter begins to reveal the gifts that become lifelong companions.
An education in the humanities doesn’t make us better citizens or stronger coders. The humanities furnish the mind with stories that move us, art and music that inspire us, and informed questions that probe the fundamental nature of mortality, love, sacrifice, and evil. They provide, in Ezra Pound’s words, “the light” that “sings eternal.”[5]
That may not be enough to rescue an education in the humanities from its current demise, but it surely should be.
[1] By 1996, 17-year-olds were already reading for fun much less often than in 1984. National Center for Education Statistics. (1996). NAEP 1996 trends in academic progress. U.S. Department of Education. The decline in the number of students enrolling in the humanities in the US started in the 1970s.
[2] In this essay, I focus on the humanities in particular, not the “liberal arts” which, for centuries, have also included math and some science (in the quadrivium). But arguments for the liberal arts today focus on the value of the humanities, and today most universities separate the arts and sciences. In the Forbes piece I cite, the author writes of the liberal arts: “At the level of a college major, think of degrees in fields like English, history, and philosophy”—in other words, the humanities.
[3] See for example, Martha Nussbaum, “Not for Profit: Why Democracy Needs the Humanities” (Princeton University Press, 2010).
[4] Peter de Bolla, Art Matters (Harvard University Press, 2003), p.53.
