1.
In her latest viral essay—its virality owing largely to the wrong reasons—Canadian-American journalist Vauhini Vara calls attention to AI’s gradual and continuous infiltration of literature. Published in The Atlantic, “Yet Another Literary Prizewinner Seems to Be AI” accuses this year’s Caine Prize winner Simbarashe Steyn Kundizeza of writing his story “The Blue Fighter” with the help of a large language model. Towards the last movement of her essay, she hesitates a moment to admit the possibility of being wrong. “But if I’m right,” she writes, “it means that several of the world’s most prestigious literary prizes have been infected by language that represents not people’s actual values and cultures, but a damaging mimicry of them by products of the most powerful companies and people in the world.” And “this is the time to do something about it.”
Still, Vara’s damning case is built completely on the word of a machine—Pangram, an AI detection software she insists its developers claim is not only accurate, but overwhelmingly so. There are, however, other telltale signs, she adds. “What I also noticed, more disturbingly, was the presence of stereotypes about African life common among outsiders: a blundering, unemployed father; a resilient matriarch who dispenses clichéd approximations of wisdom; an environment of deprivation.”
Vauhini Vara is not unfamiliar with modern technology. After schooling in the backyard of America’s hotbelly of techno-commerce, in Stanford, she’s now closing in on a two-decade career writing about the going-ons in that apocalyptic minefield. In 2024, after a two years prior critically-received debut novel, she published a book she wrote with the help of a chatbot. She’s traded in both worlds, writing by the senses the creator gave her and by the ones strong-headed capitalists are mendaciously taking up a quarter of our planet’s energy and water for. It is not, therefore, out of place for her to lay claim to knowing one or two about all that. But then one wonders what this bloody Canadian knows about Zimbabwean culture, especially after she gets the very first thing wrong, talking about African tropes, as though that beautiful country alone—never mind Africa—with its dreamy meadows and gentle valleys does not contain more ethnic groups and cultures than she can name.
2.
One wonders about Vara’s motive. Although, to be fair, she’s always, in previous writing, worried about artificial intelligence. How it is owned by a powerful few who become even more powerful by so doing, hijacking our lives as we know it, especially the little remaining grains of agency we have, now depleting day by day—language. Her tone in another essay “What If Readers Like AI – Generated Fiction?” bordered on alarm. In it, she reveals her findings of a fine-tuned AI model which demonstrates frightening potential to not only deceive readers through perfect mimicry of established writers but also producing even better, more visceral prose than theirs. What’s more? Readers—and not only average readers; for even she—appeared in a blind test to prefer the machine-generated fiction. (I remember reading the essay and sharing those fears.)
“This suggested that anyone with some storytelling skills could feed a plot into a fine-tuned chatbot, slap their name on the resulting manuscript, and try to publish it,” Vara wrote. And, yes, “—after all, we read books to access someone else’s consciousness. But what if we can’t tell the difference?” Fears of a future where AI-generated literature is normalised. Fears of a Sam Altman-manipulated future spawning a new dominant concept of the novel; one entirely devoid of human connection. Fears, valid fears. And it’s easy to see that Vara’s call to arms, expressed in the same essay, to “use our literature—novels, but also the other forms of communication available to us in the digital age—to collectively envision and enact all sorts of alternatives to the future Altman describes (‘a new way to interact with a cluster of ideas that is better than a book for most things’)” is not only noble, but urgent.
Still, one gets the sense that Vara does not really believe artificial intelligence could excel at language. In her essay “Humans vs. Bots—Who Does the Em Dash Better?” referring to Julia Marquez-Uppman’s essay, “An Abrupt Break in Thought,” she wrote: “The unexpected uses of the em dash in that essay bear no relation to the unfailingly trite way it appears in chatbot output.” She’d on that contested ballot unflinchingly thumb-printed for humans.
This bias appears to be a machine problem, too; understandably, because it is still humans who are behind these machines. Vara reveals that Pangram failed to flag AI-generated fiction samples from a fine-tuned AI model (citing this research by Tuhin Chakrabarty, Jane C. Ginsburg, and Paramveer Dhillon). And so, it is completely unsurprising that Vara would begin her case against “The Blue Fighter” by decrying its quality. “Taste is subjective,” she writes. “But trust me, this story is bad—maudlin and full of tired tropes.” These days, bad writing—especially writing that has won something and is perceived to be bad—gets slapped with an accusation of AI plagiarism. It is a curse of writing in the modern age. And this has so far come upon coloured writers in an overwhelming disproportionality, a ready-set cudgel in the hands of a new regimen of the language police.
3.
Vauhini Vara is all of us—a big fan of legendary Kenyan writer, Ngũgĩ wa Thiong’o. She must be. For she quotes him judiciously: once in the aforementioned New Yorker essay, to emphasise the threat of cultural incursion AI-generated language poses to our future. “Thiong’o’s approach is instructive in the age of artificial intelligence,” she writes. “His work reminds us that language generated by AI also carries culture and values—which the companies behind the models largely determine themselves.” To buttress this, she adds, tongue-in-cheek, referring to the same fine-tuned AI model which failed nonetheless in reproducing a Spanish expression of Junot Díaz correct: “I imagine that if Díaz’s imitator had a face, it would look less like him and more like Altman in a sombrero.”
Vara quotes wa Thiong’o again (same quote by the way) more recently, in her indictment of Simbarashe Steyn Kundizeza. This time, too, she is emphasising the threat of AI-generative language. But it serves two other purposes. Seeming to say, one way, that look, wa Thiong’o said “Language carries culture, and culture carries, particularly through orature and literature, the entire body of values by which we come to perceive ourselves and our place in the world,” but Kundizeza, his black kin, employs a language in his story that is completely devoid of that! Another way, Vara appears to be appealing to the literary community; to change its approach to things, as wa Thiong’o did, choosing to return to his native language of Gĩkũyũ after having built a career writing in English. And one cannot help but ask: is she referring to Simbarashe Steyn Kundizeza and other African writers as well?
Vara’s tone is haughty and dismissive. Careless, even—who she be sef? Posing as the Idi Amin Dada of global literature.
At the very least, there are nuances to consider. Each story demands its own style of the writer, and while some writers are able to mould a style they write in throughout their careers, most do not, preferring to be formless. And there’s the fact that some, if not most, African writers today grew up in cities, without the grandfather and the thatched hut and stories by the fire, where there’s only a little of that Africanness the world has known and stubbornly, for that need which Chinua Achebe talked about in “An Image of Africa,” refuses to let go. And that for many other writers disconnected from this show-your-Africanness in fiction, displacement is bound to reflect in their use of language—and what if some writers simply choose to write this way?
But Vara clearly did not set out to offer criticism. Or perhaps is just not interested. She does the simple thing, however, arriving with judgement which isn’t above the illiterate summation of idling market women on a slow business day: “Trust me, this story is bad.” Then invokes the AI gods of Pangram. No, we do not trust you, Vara. As a matter of fact, we barely know you.
To invoke Karl Popper’s falsifiability theory, Pangram, or any AI detection software for that matter, cannot in a scientific sense be accepted as proof of AI usage, for all Vara’s advertising of it. It is merely a highly corroborated hypothesis, which by its very design is bound to fail many times. It already has and continues to. And there’s something else: it has shown to be biased against non-native English writers. Thumbs to Caine Prize’s trustee chairman, Ellah Wakatama, who says, as reported by Vara: “My stance, as the person who leads the prize, is that if there is such an accusation, I want proof that does not come from a machine.” And that “she plan[s] to approach Kundizeza with compassion and the presumption of honesty.” It leaves serious questions as to why this wasn’t Vara’s approach to begin with.
4.
There’s much to be said about language and writing, especially as it pertains to the African. There have been debates. How should the African writer use language? What language must he write in? What must he write about? And this and that. It is already well-known—it weakens the limps of dance, beating the new drums to the tune of an old song—the Makerere African Writers Conference of 1962 and its famous resolution of non-resolution. Soyinka, Achebe, and Ngũgĩ all taking differing positions. And its implication being that the African writer, just like any writer anywhere in the world, is at liberty to write whatever and however the hell he chooses. An old song sung long ago, before Vauhini Vara learnt how to spell her name.
I read the Caine Prize shortlisted stories when they were announced in September. I recall remarking to a writer friend how a story of his was stylistically similar to Kundizeza’s “The Blue Fighter.” He had sent me a draft to review a month prior. A spontaneous short story that sprouted from the fertility of a novel-in-progress. A nice little story. A profoundly sad story. And I told him that though he’d made better use of the child narrator than Kundizeza, the styles were similar. The detached, flatness of tone and dialogue that is a botched, “clichéd approximations of wisdom” (I agree with Vara on this). However, I, as well as three other writers, can vouch that no help from any LLM was sought in the wrought of that story. We saw it develop right in front of us. But if Pangram flagged “The Blue Fighter,” it may as well flag that story too, you see.
It is not as if some critics on the African continent do not squirm at this kind of prose too. More or less what Carl Terver calls “just-there” prose, the “flattened, tired, bromidic forms of prose gradually evolving into slop,” and wonders in a craft essay, “Finding Tsav Fiction and the Tiv Prose Style” (unpublished), what point it is in “writing fiction that isn’t different?” Chimezie Chika has also once cried, “Why have we not produced the kind of prose stylists that define generations?”
Whether readers love this prose or not is entirely a different matter; I guess this is why “The Blue Fighter” won. Left to me, it’s a no. It wasn’t the favourite of many critics either. Chika aptly remarks in his review of the shortlist: “There is no reason why a contemplative story like this wouldn’t have read better, but it doesn’t. […] What sticks out most here is the dialogue. It has a certain artifice worsened by a propensity toward aphorisms.” And although reluctant to make a prediction, he wagered his “inflated naira” on Akinwande Jordan’s “So Long It’s Starting To Look Like Eternity” or Yasmine Zohdi’s “The Absence of Stain.”
Even Immanuel Fabian who published his first critical essay, on the Caine shortlist—and perhaps because of this galloped around the page with his ink, handing out colourful balloons to all participants, his tone generous and cheerleader-like—had nothing special to say about Kundizeza. Another reviewer, Felix Eshiet, accuses the Caine Prize of favouring gynocentric literature in this year’s edition:
A prize of Caine’s timbre ought to want more than a strong collection of variations on one subject and everything it spectrums. It ought to want collision. It ought to want surprise. And above all, it ought to leave us with the sense that African fiction is panoramic. […] Simbarashe Steyn Kundizeza tells “The Blue Fighter,” a pedestrian plot, in serviceable, if not, middling aphoristic prose that somehow manoeuvres a decent narrative pull, but in a child narrator POV that not only emotionally over-signposts but struggles to attain naturalness.
As for his favourites to clinch the $13,000 in London, he mentions Melara Mvogdobo’s “Dieu Merci” and Gladwell Pamba’s “The Last Shop.”
The incontestable truth, however, is that taste, as Vauhini Vara herself admits, is subjective. And perhaps she should have stopped there, instead of letting her obsession with AI and her fanatical belief in fallible software—coupled with what Wole Soyinka calls the beholder’s cataract—get the better of her in wrongly diagnosing a man with leprosy and declare he has it (Kundizeza has denied the allegation and I believe him), when he may simply have a mild case of bad writing.♦
Related
Proof of Life: On Being Accused of Writing with AI
Yet Another AI Birdoplane: On Vauhini Vara’s AI Writing Hunts
Subscribe to Afapinen to read more insightful essays.
Kenneth Kumater Katsina has published a short story “Séance,” and is completing a debut novel.
