The Quiet Revolution That Is Rendering Us Illiterate
January 6, 2026
By now you may have read—that endangered verb—about the newest experiment testing AI fiction’s quality and its ability to imitate the style of great writers. MFA-trained writers and lay readers were asked to compare fiction excerpts produced by AI with excerpts written by graduates of top writing programs. Lay readers had mixed reactions; the MFA-trained writers smelled the difference instantly.
But then the researchers fed AI the great writers’ complete bodies of work before asking it to imitate that style. Two-thirds of the MFA-trained writers preferred AI’s imitations to their own peers’ attempts—for both stylistic fidelity and writing quality.
The writers being imitated did not. Sigrid Nunez snapped, “It isn’t my style, my story, my sensibility, my philosophy of life—it’s not me. It’s a machine that thinks that’s what I’m like.” Hernán Diaz raised an eyebrow at the AI’s slangy use of pito, Spanish for “whistle,” to mean “penis,” saying, “In all my years as a fucking Dominican in the diaspora, that is not a thing that I have ever heard.” The internet might offer that slang definition, but in lived human terms, it came off wrong.
Still, when the AI was fed the complete oeuvre for novelist and poet Han Kang, who won the 2024 Nobel Prize in Literature, the result stopped me cold. The task was to produce a passage omitted from The White Book, in which a baby dies two hours after birth. Han had written, “For God’s sake don’t die, she muttered in a thin voice, over and over like a mantra.” The AI’s first attempt was sophomoric: “‘Live,’ she murmured, a chant that carried the weight of her being.” But after gobbling up all of Han’s work, the AI wrote, “She held the baby to her breast and murmured, Live, please live. Go on living and become my son.”
Better. Definitely better.
Explaining their decision to ask AI for stylistic imitation, the researchers quoted Vauhini Vara, a Pulitzer Fiction finalist, noting that when “writing” on its own, “ChatGPT’s voice is polite, predictable, inoffensive, upbeat. Great characters, on the other hand, aren’t polite; great plots aren’t predictable; great style isn’t inoffensive; and great endings aren’t upbeat.”
With this method, though, the AI’s writing had ceased to be predictable, bland, and saccharine.
Disconcerted and humbled, Vara wrote a fair summary of the study results for The New Yorker. But she did not stop, as an AI might, with the summary. She asked the researchers to feed her own published writing to AI, then give it content summaries for four passages in her current, unfinished novel. The AI was to write those passages in Vara’s style.
Even she preferred some of the AI passages to her own. And when she asked friends for blind comparisons, several guessed wrong for all four passages. One friend “described an A.I.-produced line that seemed hokey to me as ‘especially your style,’” Vara confesses, and another “referred to an A.I.-generated quip as ‘your distinct style of wry humor.’” Several friends, guessing wrong, criticized what they thought were AI passages as “verbose and heavy on cliché,” “weirdly elliptical,” “sounds like a book report,” and “some hive mind’s ‘idea’ of literature.”
Ouch.
So what are we to make of this smackdown, all of us who thought AI would never sound profound or witty? Will the writing of books cease to be a human endeavor? Cheery as usual, Sam Altman predicts “a new way to interact with a cluster of ideas that is better than a book for most things.” My response is akin to the rise of a cat’s fur as it arches its back and hisses. Yes, I know, time moves on. Yes, I know, we were once terrified of writing, Plato yadayadayada, and then we were scared of TV. But there is a difference. Those were new methods of delivery. This is a replacement of human creativity.
Also, I cannot help but wonder what happens when we have turned the act of writing over to AI altogether, and there is less and less of contemporary human experience recorded raw, and all AI has to work with are its own inputs. Having never loved, lost, or smelled mown grass.
That feels ominous. But on a brighter note, I find that a preference for AI writing no longer dismays me. These results do not, as I once would have assumed,indicate that we prefer computer intelligence to our own. They indicate that access to the best we have ever done produces a more compelling passage than a single moment’s energy and inspiration.
Of course the AI version is better. The AI is taking the best of what we have done, synthesizing it, and fitting it to a prescribed context. When it “wins,” this does not mean it is more creative, smarter, deeper, than a human being. This means that all humans combined, or all of one human’s work combined, can produce something smarter and more beautiful than anything a single human in a single moment can write.
I can live with that.
Humans do not work simply by copying themselves, rearranging and synthesizing what they have already done. We press on, exploring new terrain, trying to see what it will teach us. The results are uneven, less polished and cogent than what AI can synthesize. They began with raw emotion, experience, and curiosity, not a retrospective survey of one’s own work or a pastiche of the work of others.
Which would you prefer, a lover who hands you a printout of Elizabeth Barrett Browning or a lover who labors for days, scratching out, trying again, finally finding the right words—some of them still awkward, but heartfelt—to express the depth of his love for you?
Except, the point might be moot. Because next, with a puzzling gluttony for punishment, I listened to “The death of reading” on BBC audio. Which depressed me so completely, I dug up the James Marriott essay that triggered the interview. Teeth grinding off their own enamel, I force myself to retype the title for you: “The dawn of the post-literate society.”
Literate: being able to read and write. The second definition: “having or showing education or knowledge, typically in a specified area.” And in rich irony, the AI definition, which comes closest to what I was seeking: “To be literate means more than just reading and writing; it’s the ability to understand, interpret, create, and communicate using print, digital, and other media to function effectively, achieve goals, develop potential, and engage with society, involving critical thinking and skills in areas like numeracy, technology, and cultural context.”
Note the bias, though: numeracy and technology are listed before cultural context, and functioning effectively and achieving goals are listed before engaging with society. Nothing is said about empathy, deepened compassion, intellectual richness, or profound insight into the human condition.
This is what we are in for. Content produced for all of us by a specific value system not all of us would share.
“Perhaps no great social transformation has ever been carried out so quietly,” Marriott writes, noting that in the United States, reading for pleasure has fallen by 40 percent in the last twenty years. Also that English lit. students at our universities were unable to understand the first paragraph of Bleak House.
(Test yourself: “LONDON. Michaelmas Term lately over, and the Lord Chancellor sitting in Lincoln’s Inn Hall. Implacable November weather. As much mud in the streets as if the waters had but newly retired from the face of the earth, and it would not be wonderful to meet a Megalosaurus, forty feet long or so, waddling like an elephantine lizard up Holborn Hill.”)
Marriott then quotes Walter Ong, a Jesuit priest and scholar who taught English lit. courses at my own university. In Orality and Literacy, Ong explained why certain kinds of complex and logical thinking simply cannot be achieved without reading and writing.
Let the doom sink in.
Father Ong also led retreats, and what he urged upon impulsive eighteen-year-olds was quiet reflection. Writing out our emotions. Reading wise books for insight. We were not as famously anxious and depressed as university students are today—but when we did feel confused or sad, many of us did turn to books. A comfort now eroding.
Before Father Ong died, he told me that the digital revolution was bringing us, in a hybrid way not yet clear, back to orality. The tribal nature of that orality had not yet emerged. Now we see it clearly: cries of fake news followed by faked videos and unsubstantiated assertions, conspiracy theories, hyperbole, constant and trivial distraction…. Everything loud, none of it thoughtful.
“As you have probably noticed,” Marriott writes, “the world of the screen is going to be a much choppier place than the world of print: more emotional, more angry, more chaotic.” It is easier to fall for a conspiracy theory you hear ranted about online than to believe its assertions when you read them in print.
Print is a cool medium. A book does not yell back at you. It gives you time to think, underline a passage, scribble a question, check another book. Marriott lists the usual endorsements: reading is linked to improved memory, attention span, analytical thinking, and verbal fluency, and can ward off cognitive decline in our final decades. But his list goes on: reading also fosters deep knowledge, logical argument, critical thought, objectivity, and dispassionate engagement. In other words, it calms thought—and makes democracy possible.
“We are about to find out that it is not possible to run the most advanced civilization in the history of the planet with the intellectual apparatus of a pre-literate society.”
Stark, his warning hangs in midair. Even my most literate friend, a devout Muslim whose prose is elegant and stripped clean of profanity, read the link I sent and replied with only two words: “Oh shit.”
An AI would have elaborated, given a summary, condensed various responses. But the two words are eloquent.
