The Damned Mob of Scribbling Women
Amid rising fears of AI-driven creative production, Brianna Di Monda interrogates the explosion of fanfiction in mainstream markets, and what it might tell us about the idea of literary value today.
A virgin maiden abandons her ceremonial role to hunt vampires with her immortal bodyguard. A woman falls for her faerie kidnapper and goes on a quest to save his kingdom. A war college misfit has eyes for the son of a rival family—and their dragons are mates. The most popular romances today aren’t simple love stories. And as absurd as the plots come, these books are mainstream: romantasy, a term popularized in the last five years to describe the longstanding blend of fantasy and romance, is drawing tens of millions of readers. In 2023, Amazon offered millions of romantasy e-books, up from 600,000 in 2014. The #romantasy hashtag has over 800 million views on TikTok. The books spend months on New York Times bestseller lists, and the Goodreads Choice Awards gave romantasy its own category in 2023 to stop the books from competing with traditional fantasy novels. In the last two years, the number of romance-only bookstores across the country grew from two to more than twenty. In 2020, 18 million print romance novels sold; in 2022, romance sales in the US shot up by 52.4%, compared with an increase of just 8.5% for adult fiction overall. 2023 saw the sale of more than 39 million romance novels, and in 2024 the Oxford English Dictionary shortlisted romantasy as their Word of the Year.
Romance is a multi-billion-dollar genre that has been said to account for nearly 40% of all book sales in the last decade. Publishing has never had a bigger mass market. Since the ’70s, there’s been talk about finding new paths for the novel that express and cater to women’s desires. For better or worse, we’ve come to a new form of feminine writing, and it’s romantasy.
There have been grumblings about the quality of these novels: Tank Mag predicts A.I. will replace romance writers entirely, dismissing them as the “machinic face of automated creativity.” The New Yorker did a podcast on romantasy, where one critic said the genre is “not even literature, it’s A.I. It’s exactly made for A.I. This is what A.I. should be for and just this.” Then there’s The Baffler's piece on dark romance, which acknowledges the appeal of these stories and then calls the genre one that “disempowers women, naturalizes abuse, and continues the false belief that complex sexual desires are only harbored by harmful and irredeemably aberrant people.” Mark McGurl, the author of Everything and Less: The Novel in the Age of Amazon, wrote of the romance novel, “Surely a badness so profound, so completely disqualifying of critical esteem, or even sustained curiosity, has something interesting to tell us about the contemporary literary field as a whole?”
[book-strip index="1"]Anyone who reads romance will tell you that they’re not difficult books. This is the genre’s strength: accessible language means readers don’t get bogged down by the writing. Perhaps this is what George Eliot meant when she criticized romance novels in 1856, calling them “the most trashy and rotten kind of feminine literature,” written by women “inexperienced in every form of poverty except poverty of brains.” Literary fiction is in theory a more noble pursuit: at its best, it has an expectation of moral betterment (“style is morality” à la Martin Amis). It emphasizes subtle, descriptive, and often flowery language. The focus is on how the sentence appears and its psychic resonance, not on being immediately legible. For Eliot, who considered romance novels “remarkably unanimous in their choice of diction,” their lack of emphasis on language was “pitiable.”
There are two juggernauts of romantasy: Sarah J. Maas (of A Court of Thorns and Roses, or “ACOTAR,” fame) and Rebecca Yarros (Fourth Wing). The books of Maas and Yarros have grossed more than half a billion dollars—without a television or film adaptation. In 2023 alone, the sales of these two authors rose 45% to nearly 20 million copies, while the combined print sales of James Patterson, Stephen King, and John Grisham came to six million. To placate Eliot’s ghost, I’ll admit: their writing is desperately lacking, as is true in most mass market books, in style—precisely what makes romance seem like a prime target for A.I. generation. At the same time, and poor structure can make the series a slog. ACOTAR is rife with clichés and scenarios out of sexual fantasies. No one could possibly mistake a Rebecca Yarros sentence for one by Donna Tartt.
But not every book needs to demonstrate a writer’s ability to handle a sentence; this is simply not always the implied contract between reader and author. Genre readers seek out character development, worldbuilding, quick plot progression, and familiar tropes. They want a compelling, if predictable, story. It’s not just romance readers. Frederic Jameson, in Magical Narrative: Romance as Genre, says all genres are essentially “contracts between a writer and his readers.” Or, to quote McGurl again: “all fiction is genre fiction in that it caters to a generic desire.” He equates attachment to one genre with a child’s “demand to be read a favorite story again and again.” From this perspective, literary fiction isn’t above the dismissal of genre as a form of therapy or self-care. It’s just another genre, catering to a generic desire.
[book-strip index="2"]An A.I. cannot enter into this contract; it writes without reference to meaning. A.I.-generated romance books are already appearing on Amazon, and consumers demand their takedown (in symbolic recognition of the problem, Amazon now limits Kindle authors to publishing three books a day). A machine can’t write from lived experience, and this is a huge drawback for readers in a genre based on shared experience, healing, and community. They already know that our own life is all that separates us from machines, and if we give creative work to A.I., we kill the root of our shared impulse to form, reform, and advance stories through collective participation.
We develop different attachments to characters in literary fiction and fantasy. Identification with a fantasy character is seen as cringe (cosplay, midnight showings, larping), while reading literary fiction grants access to what Roland Barthes calls “the effect of reality.” We form parasocial attachments to the Great Authors, elevating their characters to archetypes of the human experience. Romance, as first defined by Northrop Frye, is a wish fulfilment: it is “the search of the libido or desiring self for a fulfilment that will deliver it from the anxieties of reality but will still contain that reality.” Literary fiction is meant to tell a story that is pure, intelligent, and true, to lie with its readers in bed and assure them that their existence is already meaningful.
My friends who read literary fiction never touch genre; my friends who read romantasy tend to only read genre fiction. There is an earnestness to women’s consumption of romance, and the women I talk to gush about Maas’ ability to write the aftermath of trauma. The story involves manipulative relationships, abusive siblings, coercive sex, and queer children sold as brides. In literary fiction today, women tend to be amoral, frank, bleakly funny, very smart, and perverse in their motivations in ways that destabilize the reader’s assumptions about what is ugly, what is desirable, and what is permissible. But romance novels go further: they embrace a broad spectrum of female fantasy, from progressive narratives featuring women who maintain their independence, to traditionally “unfeminist” desires of domesticity, to manipulative power dynamics with possessive or controlling heroes. These books recognize an expanded liberation that allows readers to explore socially unacceptable desires. By unabashedly writing these stories, romance novels create a space for women to claim complete psychological freedom and sovereignty over their imaginations. While critics focus on the normative gender roles, there’s another, more obvious, difference in genre: while the trend in much recent literary fiction is to tell stories of suffering and abuse—the popularity of “sad girl lit” comes to mind—romance is a female-dominated literary space that offers happy endings.
It can get tiring to read contemporary literary fiction, to be left with dozens of harrowing stories and little resolution. ACOTAR does not quite make darkness visible; it uses these elements to give characters backstory and motivation, and put the energy of the plot into how they move on from the past. As Vulture wrote, “At a cultural moment when the trauma plot feels so familiar, it’s no wonder a series that treats it as a lingua franca has become so widely read. Fantasy has a surfeit of Chosen One narratives; in Maas’s work, the Chosen One needs to learn to Choose Herself.” There’s even a library built into a mountain that employs and protects survivors of sexual assault. Romance’s commitment to emotional resolution creates a deeply human exchange that A.I., for all its pattern recognition, cannot authentically replicate—though critics continue to frame romance as formulaic enough for machines to master.
[book-strip index="3"]The 2015 documentary Love Between the Covers argues that the romance genre has the most diverse readership—at least across age, sexuality, class, and race—and brings thousands of women together. The authors form real relationships with their readers, often successfully encouraging them to become writers themselves. Similarly, Janice Radway, after researching the popularity of romance, writes, “Romance reading, it would seem, profoundly changes at least some women by moving them to act and to speak in a public forum.” The romance community considers itself one of the last meritocracies in publishing, where reader enthusiasm remains the most important factor in determining which authors gain prominence.
Domestic fiction experienced a boom in the mid-nineteenth century, when the most popular authors were all women (including Jane Austen). Women read these “dime novels” throughout the day, consuming, much like contemporary romance readers, multiple novels a week. Writing a year before Eliot, Nathaniel Hawthorne was furious at the genre’s sales, saying, “America is now wholly given over to a damned mob of scribbling women,” beginning a long tradition of critical dismissal.
Pamela Regis, in A Natural History of the Romance Novel, ties the genre’s origins to the rise of affective individualism, where the individual’s fulfillment in love became the point of life itself. The romance novel expressed these new social values; not only did it become taken for granted that people married for love, but a woman’s marriage became all-important because she lost her property rights and could rarely obtain a divorce. “No wonder a woman’s choice of a spouse was so fraught with drama,” Regis writes. Romance was the first genre to treat women’s sexual desires positively and reward their agency. Until romance, literature was filled with dead or tortured women, the two most notable exceptions being the Greek virgins Athena and Artemis.
At issue in the romance, in Regis’ words, is not the marriage, but a heroine moving “from a state of bondage or constraint to a state of freedom.” The realm a romance inhabits is an “ideal, non-mimetic, fantasy world.” The question of reality in the romance genre is not something that has ever been the point: even titles like Pride and Prejudice, Jane Eyre, or A Room With a View have unrealistically happy endings, but nonetheless capture the imagination. Fantasy is a powerful thing, especially in relation to desire and love. Romance has always been about a way to be in the world otherwise. This was true especially when the split between romance and romantasy started to occur: in 1976, Anne Rice’s Interview with the Vampire bridged gothic romance and fantasy, though it wasn’t marketed as romance. Still, The New Republic gave it a scathing review, saying the book had no purpose beyond its “suckling eroticism.” Yet the book’s success—it is one of the best-selling novels of all time, with twelve sequels and a movie adaptation—demonstrated the growing acceptance of supernatural elements in mainstream fiction and the appeal of turning fantasy creatures into love interests. It also gave rise to a slew of vampire romances like Linda Lael Miller’s Forever and the Night (1993), Amanda Ashley’s Embrace the Night (1995), Ronda Thompson’s My Darkness (1997), and Nina Bangs’ Eternal Pleasure (1997), to name a few. All of these worked to establish a new set of romance conventions like vampire aristocracy, fated mates, and immortal love—tropes that would take off with the rise of the internet.
[book-strip index="4"]Today’s romantasy boom grew out of Harry Potter fanfiction, an online community that aughts readers flocked to as they waited for Rowling’s next book to come out. Anne Jamison’s Fic: Why Fanfiction is Taking Over the World notes that in 2001, “studio attorneys sent cease-and-desist orders to operators of several fan sites in the United States and Britain.” By 2004, studios changed tactics, courting these communities as the Harry Potter frenzy became “directly related to the growth of the Internet” and its fanfiction phenomenon. Even the epilogue of the seventh book—which teases the personalities and love interests of the trio’s children as they board the Hogwarts express—is bizarre to read today without the context of fanfiction’s prominence: hundreds of thousands of websites were dedicated to its writing, each with the ability to keep the fandom alive long after the series ended. Around the same time, Stephenie Meyer’s Twilight saw explosive success in 2005, in turn establishing the YA paranormal romance category beyond fanfiction forums and further cementing the foundations of romantasy.
Several of these fanfiction writers got book deals, two of the most notable being Cassandra Claire and Sarah Reese Brennan. Maas herself started by writing Harry Potter fanfiction. Archive of Our Own (AO3), launched in 2008, revolutionized fanfiction writing by allowing users to tag fics by tropes (some of the most common being “Age Difference,” “Possessive Behavior,” and “Unrequited Love”). This changed how readers sorted and discovered new stories. It introduced a “folksonomy” system where users could create their own tags for their stories, which volunteers then standardized and listed across fandoms. Although fanfiction started as an alternative to commercial platforms, which these readers saw as not catering to their interests, publishers began to study the tagging system to understand market trends, and books were advertised on Goodreads and Amazon according to popular fanfiction tags. Most fanfiction was “steamy,” reflecting readers’ desire for more mature YA fantasy—the most popular Potter fanfiction involves Harry/Draco pairings, or “slash fics.” This pointed to a significant oversight in traditional publishing that persists to this day: while queer narratives dominate online fan spaces (from 2008–2021, a little over 50% of relationships on AO3 were same-sex), traditionally published romance has been slower to reflect this diversity. The anonymity of fan platforms allowed writers to explore narratives without commercial pressure or personal exposure, while publishers continued selecting more commercially “safe” narratives for mainstream publication.
Harry Potter fanfiction continues to get writers book deals: SenLin Yu’s Manacled, published on AO3 between 2018 and 2019, has nearly sixteen million individual downloads and is slated to come out as a standalone book with Penguin Random House in September. The fanfiction-to-book pipeline is so common that the community calls it “filing off the serial numbers” when their fanfiction gets enough traction to secure a book deal and they remove all infringing IP. Fifty Shades of Grey, published in 2011, propelled the rise of fanfiction by revealing its vast untapped market. In 2023 Vulture covered the growing number of fanfiction writers in publishing lists: “The appeal is understandable: Fic writers bring knowledge of how to market a story and build an audience, a boon for editorial houses. The fans authors have gained writing fic will buy books, in some cases carrying them to the best-seller list.” Had Stephenie Meyer chosen to sue E. L. James for copyright—and she had a strong case, as James only swapped out the character names before publishing Fifty Shades—she could have kneecapped much of the market and fanfiction communities that have emerged in the last fifteen years.
Beyond their complicated taxonomy system, AO3 also created a documented history of how tropes have evolved and combined, tracking the minute changes in reader interests. There’s been trope hybridization as “Fated Mates” blended with “Enemies to Lovers” (which itself had previously merged with “Arranged Marriage” and “Forced Proximity”), a story that became a cornerstone of romantasy. There’s also been a modernization of traditional romance plotlines: the “Marriage of Convenience” evolved into today’s “Fake Dating.” AO3 has also witnessed the reader preferences becoming incredibly granular: “Slow Burn” evolved to include the specific word count expectation of 100k+, “Mutual Pining” split into subcategories (“Only One Knows,” “Both Know But Can't Act”), and “Grumpy/Sunshine” emerged as a distinct category from the more general “Opposites Attract.” Meanwhile, “trope subversion” became a distinct category of writing.
Frederic Jameson wrote that romance gets “secularized and renewed” for each era’s needs: “A history of romance as a mode becomes possible, in other words, when we project it as a history of the various codes which, in the increasingly secularized and rational world that emerges from the collapse of feudalism, are called upon to assume the literary function of those older codes which have now become so many dead languages.” Romance has recognizable patterns that have come to be seen as formulaic, but the genre is constantly evolving through this historical regrounding. With AO3, this regrounding can happen quite quickly, determined by the interests of readers and writers. A machine can’t perform the cultural work of regrounding.
A.I. has effectively made it so anything plot-based can be plucked out of the Tower of Babel. Since it is trained on existing writing—being plagiarism without a clear source—it’s only the furthest extension of “filing off the serial numbers.” To paraphrase the definition given by Matteo Pasquinelli, A.I. is “a statistical tool of multidimensional analysis” applied to “the automation of pattern recognition.” Romance similarly iterates plots based on a shared language model of fanfiction and tropes. There’s already an employment of ghostwriters and collective pseudonyms, as well as a heavy reliance on tropes that plagues the lesser books in the genre. There’s a bittersweet irony that the writers for a genre about articulating positive desire are thought to be most at risk for being replaced by robots. But this overlooks the difficulties other genres have faced: the science-fiction magazine Clarkesworld cut submissions after an onslaught of A.I.-generated stories, and Asimov’s Science Fiction now requires authors to verify their work is not A.I.-generated.
Before the threat of A.I., there were already repercussions for romance’s inherent flatness. There was the infamous Omegaverse lawsuit, where an author tried to trademark an entire subgenre. Others have tried to trademark generic phrases like “dragon slayer” or the words “dark” and “cocky.” Now, one romantasy author is suing another for copyright infringement on a “Vampires in Alaska” story. But you can’t trademark—or claim plagiarism on—a genre’s tropes. One of the defendants in the vampires case called the lawsuit a threat to the creative commons.
[book-strip index="5"]“The most thought-provoking thing about our thought-provoking age is that we’re still not thinking,” Heidegger writes. His thesis is that, though writing is central to our pedagogy, thought has nothing to do with language, and everything to do with being in the world. As Markus Gabriel argues, thinking is embodied: it’s a sixth sense, and a machine cannot think because it doesn't have a bodily relationship with the world. It doesn’t have an internal dialogue. The snobbish satisfaction that A.I. could take over romance focuses not on the content or reception of A.I.-generated work, but that a machine could seem to express the thoughts of a person. It doesn’t do it so well, which is why it will come for this “lesser genre.” Perhaps I can appropriate something Ted Chiang said: “Generative A.I. appeals to people who think they can express themselves in a medium without actually working in that medium.” Critics are derogatory about A.I.’s ability to take over romance because they think the genre is highly replicable, that it doesn’t involve higher-order thinking, but they don’t engage with how or why women interact with the genre. It ignores, too, that even art made for entertainment requires fine-tuning; it’s not so easy to write a bestselling romance novel.
No one has demonstrated a desire for books created by A.I.; even Google’s Paris Olympics ad for Gemini—in which a young girl composed a letter with A.I.—was met with widespread backlash. This is because even the writers whose sensibilities offend us the most understand that the effort involved can’t be divorced from the making, and that iteration cannot be equated with invention. This raises interesting existential questions of what writing is, and it may be that, even if a machine can technically write a novel, style never mattered so much anyway. The throughline that protects artists is, in the end, art’s middlebrow earnestness: that we want stories to be reflections of a person’s interiority, to come as a result of extended consideration of one idea, and to use those stories to better understand ourselves.
Surveys conducted by the National Endowment for the Arts show Americans are reading fewer books than they did a decade ago, while romance demonstrates remarkable vitality through its participatory nature. Readers and writers convene at the Romance Writers of America conferences, at midnight release parties, on AO3, on BookTok. This mass participation by readers and writers—in particular readers that regularly become writers—runs contrary to dire predictions about the death of reading. It’s also the genre’s strongest defense against A.I. replacement, and when, eventually, A.I. gets “intelligent” enough to write a decent literary fiction book, it’s these genre communities that will have been on the frontlines, and who are demonstrating, repeatedly, that they don’t want A.I. art: they want, to use a Walter Benjamin term, aura. They know that people line up to see Michelangelo’s David even though a replica is in its original spot in the Piazza della Signoria in Florence, that “the existence of the work of art with reference to its aura is never entirely separated from its ritual function,” that people would rather have the ashes of the Mona Lisa than a perfect replica. They line up at book signings to meet this author, to shake the hand of someone who put to words something they could not articulate. They need this physical presence: the center and locus of meaning, from which all significance emanates.
--
Brianna Di Monda is the editor-in-chief of the Cleveland Review of Books.
