In the late aughts and early 2010s, before Cambridge Analytica and Theranos and the so-called techlash; before the interest rate hikes and layoffs and Sam Bankman-Fried; before Silicon Valley’s leaders pledged fealty to Trump and military contracting came back into vogue, big tech companies seemed like pretty nice places to work. There was shrimp in the office; free laundry service was a corporate perk; many enjoyed feeling that they were unique in the history of capitalism in having figured out how to do well by doing good. 

But even as droves of young coders moved West to find their fortunes, a less glamorous, mirror workforce was beginning to form. In office parks in Phoenix, Austin, Nairobi, Hyderabad, and Manila, content moderators spent full workdays reviewing flagged posts and removing the ones that violated rules imposed by platforms like Facebook and YouTube. Mostly contractors, moderators received hourly wages for working in what were effectively call centers for the most gothic horrors on the web. These near-invisible workers signed NDAs, contracted with third parties, and rarely spoke freely to the press.

From the early days of moderation, the psychic toll on the worker was understood as a necessary cost. In 2010, The New York Times reported that “a common strategy at Web sites is to have users flag questionable content, then hand off material that needs further human review to outsourcing companies that can do so at low cost.” Hemanshu Nigam, the former chief security officer at MySpace, told the paper that people signing up to do this job “have no idea that some of the despicable and illegal images they will see can haunt them for the rest of their lives.” Still, the field continued to grow. By 2014, Nigam estimated in Adrian Chen’s landmark Wired investigation that over a hundred thousand people around the world were doing content moderation work — both the straightforward, though unpleasant, tasks of taking down porn and violent content and the more complex ones of evaluating hate speech, satire, and bullying. 

As online mischief became more sophisticated and systematic in the mid-2010s, and political norms evolved, the tech companies quickly adapted their guidelines. After the 2016 election, as part of a general effort to show they were doing something, anything, to clean up the sites that were widely blamed for elevating Donald Trump, executives vowed to adopt new policies, invested in fact-checkers to spot hoaxes, and increased the number of moderators. In 2018, Facebook announced that it would be roughly tripling its safety and security team to some thirty thousand people in light of “the enormity of this challenge and the responsibility we have to get it right.” 

Attitudes toward moderation broke on party lines. “Social Media is totally discriminating against Republican/Conservative voices,” Trump tweeted in 2018, shortly after several platforms banned the Sandy Hook conspiracy theorist and right-wing radio host Alex Jones. A 2019 poll from YouGov found that 58 percent of conservatives thought that removing or deleting comments from social media sites suppressed free speech, while only 29 percent of liberals felt that way. When Facebook and Twitter banned the former and future president after January 6, MAGA loyalists decried censorship — Donald Trump Jr., for one, took the occasion to tweet that “free speech is dead & controlled by leftist overlords” — while anti-Trumpers cheered. 

But as Trump re-amassed power, companies retreated. Not long after taking over Twitter in late 2022, Elon Musk gleefully welcomed the former president, along with hate speech and sundry repellent content, back to the site. Just before Trump was inaugurated for the second time, Mark Zuckerberg — who the president-elect had suggested would “spend the rest of his life in prison” if he “steered” platforms against him in the 2024 election — dismissed moderation. It was time, the Meta CEO said in a video, “after years of having our content moderation work focus primarily on removing content,” for “getting back to our roots about giving people voice.” Last summer, The New York Times reported that YouTube, which had previously instructed its moderators to take down content including slurs and election- and vaccine-related lies, had started telling its teams to leave this material up. Meta and YouTube followed X in experimenting with “community notes,” which asked users to flag misinformation in lieu of professional moderation.

While these political battles raged on, a cohort of writers got to work transforming the content moderators’ abject, harrowing lives into abject, harrowing — yet sometimes surprisingly sexy — fictional psychological dramas. In 2020, The New Yorker ran “You Will Never Be Forgotten,” Mary South’s haunting story of a moderator who “has been at this job long enough that she remembers time not only in the usual way, by seasons and holidays, but also by the content that has most traumatized her to delete.” The protagonist of Hanna Bervoets’s novel We Had to Remove This Post, published in Dutch in 2021 and translated into English in 2022, “can’t help but suspect a certain amount of lurid fascination” when people ask about her work. The 2023 play Job, which swiftly moved to Broadway and is now being staged in multiple cities, dramatizes the therapy session of a moderator who has taken it upon herself to “extract the darkness” from the internet. The 2025 thriller film American Sweatshop follows a revenge-seeking moderator who explains that “we watch this shit so no one else has to.” In Kate Folk’s 2025 novel Sky Daddy, a moderator is “happy to be paid twenty dollars an hour to flatter a machine that would soon replace me.” Elaine Castillo’s 2025 novel Moderation follows a longtime contractor who feels that “no one was better than a content moderator at dissociation.” And the April 2026 remake of cult horror movie Faces of Death stars Barbie Ferreira as a crusading content moderator who tells new hires what to do when they see alarming content: “Don’t discuss it. Don’t Google it. Don’t even think about it, if you can help it.”

These works have been lauded for the mere fact of portrayal — for offering “provocative insights about modern life” and “social media’s toxic underbelly,” as a review of Moderation in The Washington Post put it. Kirkus praised We Had to Remove This Post as “unironically, a novel for our time” that “masterfully captures our contemporary moment.” Job, the Chicago Sun-Times wrote, “is infused with a looming, ever-intensifying dread that feels all too timely.” As these critics seem to have noticed, moderators have become irresistible to writers as an au courant synecdoche for Silicon Valley’s most sensational ills. But although the genre demonstrates a real appetite for taking down the tech industry, the authors of these works seem to lack an attendant curiosity about its mechanics. They have usually undertaken enough research — seemingly from the same shallow well of journalistic source material — to make their plots and settings plausible, but not so much that they raise new questions or advance original critiques. And all of these works diverge from reality in a near-identical way: they are populated by the alienated, introspective women already familiar to readers of contemporary fiction, placing surrogates for the authors (or their audiences) into a job often done by short-term contractors in the United States or, more commonly, for poverty wages in the Global South.

In her 1922 essay “The Novel Démeublé,” Willa Cather wrote, “If the novel is a form of imaginative art, it cannot be at the same time a vivid and brilliant form of journalism.” But some of these projects read like repackaged versions of the public-interest journalism from which their writers clearly took inspiration. A 2019 Verge investigation into American moderators by Casey Newton — rigorously reported, brutally detailed, and generous to its pseudonymous sources — seemed particularly useful to these writers. The article follows a narrow set of these workers in Phoenix and describes what they deal with day to day; Newton alerts the public to the costs of having humans laboring behind seemingly automated feeds. 

Newton reports that a worker contracting for Facebook was “allotted nine minutes per day of ‘wellness time,’” a detail that recurs in both Moderation (the protagonist “rarely took her daily nine-minute wellness break”) and Sky Daddy (“I adjourned to the wellness room for my mandatory nine-minute break”). Newton also writes that moderators “have been found having sex inside stairwells and a room reserved for lactating mothers.” Sure enough, when the bosses in We Had to Remove This Post put up a sign saying “No sex acts in and around the building,” the narrator Kayleigh thinks: “We all knew that was about the lactation room.” Only a few minutes into Faces of Death, meanwhile, the central character Margot witnesses stairwell sex. In the next scene, Charli XCX appears as a moderator who smokes weed on the office roof; no surprise, since Newton describes a contractor who “used marijuana on the job almost daily.” Likewise, it doesn’t seem coincidental that, given Newton’s line about how “workers are required to store their phones in lockers while they work,” phone lockers are used as a backdrop in that film. Linda, the narrator of Sky Daddy, similarly recounts that she “stowed my phone in a locker at the edge of the moderation floor, in accordance with the company policy; they didn’t want us taking photos of the content that crossed our queues.” The reporter’s observations are parlayed into scenes: per Newton, a worker “often discussed his belief that the Earth is flat with colleagues,” and in We Had to Remove This Post, one moderator says, “The evidence that the Earth is flat is becoming impossible to deny.”

Several authors of content moderator fiction acknowledge their debts explicitly. A “Selected Sources” section at the end of Bervoets’s We Had to Remove This Post states that “any resemblance to reality is not accidental”; she also recommends “strongly to anyone interested in learning more about this topic” a list of documentaries, books, and articles, including work by Newton, Chen, and UCLA professor and leading scholar of content moderation Sarah T. Roberts. In an acknowledgments section at the end of Moderation, Castillo notes that “several books and articles,” including Chen’s, “were of great aid in the world-building of this book.” Folk, in an interview, pointed to a book by Roberts — who herself has said the 2010 New York Times article on moderation got her interested in the field — and We Had to Remove This Post as among the sources she consulted in her research for Sky Daddy. 

Echoing reality is simple enough when it comes to sprinkling in details about an office; where it gets complex is in the thorny politics of the work being done in it. Free speech clashes made content moderation into one of the most-hated, least-understood functions of the contemporary internet. But these writers address issues of online expression in a surface-level manner. Managers are brought briefly into the frame to hand-wave about free speech: “Censorship — and you know I hate the C word — but censorship is a slippery slope,” Margot’s boss tells her in Faces of Death. “We’re not censors; we are moderators,” a supervisor reprimands in American Sweatshop. The contractors learn the semi-arbitrary rules and then dispassionately, sometimes archly, apply them. (“Can you put that up online? Not if you can see blood. If the situation is clearly comical, then yes. If there’s sadism involved, no,” the narrator Kayleigh recites in We Had to Remove This Post.) In Sky Daddy, during a scene in which Linda suggests to a superior that the terms of service are “all that matters,” he tells her she was “born to be a content moderator.” 

With politics hastily dispensed with, the authors are free to embrace the pulpy elements of moderation that seem to be their central interest — the sex, drugs, and onscreen grisliness. The opening montage of Faces of Death features gunshots; a woman caught on a tree, hanging by her underwear; another woman, topless, wielding a pool cue like a club; a man getting hit by a truck; another man getting beat up; a monkey attack; a man skidding off his motorcycle; and porn. A “rule of thumb” in her line of work, Moderation’s lead character Girlie thinks, is that “if there was a hole, a face, a girl, all three, what was the difference; someone was going to try to fuck it.” In Sky Daddy, Linda reviews the flagged comments calling a teen girl things like “cum dumpster, ugly whore,” and reflects that, “thanks to my expertise, I understood these comments were in good fun.” The unnamed protagonist of “You Will Never Be Forgotten” says she “has become a veritable sommelier of beheadings.” 

James Wood wrote of systems novels in 2001 that “‘knowing about things’ has become one of the qualifications of the contemporary novelist. Time and again novelists are praised for their wealth of obscure and far-flung social knowledge.” These writers, working some two decades later, cannot be accused of knowing too much; they don’t stuff in so much as skim from information. If autofiction’s credibility comes from the “I,” and systems novels’ credibility comes from their erudition and grasp of how the world operates, content moderation fiction’s credibility strains to come from just enough research. This is, as the reporting attests, a gruesome job. But much of the work is deadly boring and tedious, too, as is any computer job, especially a low-level one. The writers’ eagerness to recount some of the most disturbing acts imaginable makes these moderator fictions seem at times less like empathetic portrayals of suffering women on the brink than pretexts for shock. 

The exact number of content moderators is hard to confirm — tech companies rely on what the scholar Roberts calls a “logic of opacity” with regard to their shadow workforces — but it’s clear that these gigs are overwhelmingly offshored to contractors in the Global South. Chen, in his Wired article, visits Manila, writing that “moderators in the Philippines can be hired for a fraction of American wages,” and earn the equivalent of a few hundred dollars per month. Although Newton’s reporting focuses on some of the Facebook moderators at a thousand-person job site in Phoenix, he notes that this represents a sliver of the worldwide workforce. The demographically unlikely heroines of content moderator fiction, meanwhile, all live in cities like Las Vegas and San Francisco and New Orleans, operating as the type of quirky, friendless, broke-ish millennial city-dwelling women that pervade many new novels. They never reflect much on colleagues abroad. In We Had to Remove This Post, which takes place in an unnamed metro area, Kayleigh briefly thinks of her counterparts elsewhere — but only to complain that moderators in India review the easier-to-evaluate porn, while she is stuck with the posts that require her to marshal her cultural competence about satire, motive, and free expression. Her types of evaluations take much longer, she thinks, and it’s harder to get through her compulsory five hundred in a day. 

The fictional content moderators may not always be happy to be at work, but their jobs become deeply intertwined with their psyches and senses of self. Their creators often take pains to portray them as exceptionally diligent. Their lives beyond the office are generally bleak, but at work, the fictional moderators are intense, capable employees: in Moderation, Girlie maintains an accuracy score of 99.5 percent and, having spent “over a decade on the job — unheard of, for a content moderator — she showed minimal signs of acute mental distress.” Linda, of Sky Daddy, is the “top mod” in her vertical, which focuses on hate and harassment comments. In We Had to Remove This Post, Kayleigh says she was “pretty damn good.” Jane, of Job, was so efficient that she worked on a team of one: “I negotiated a raise and now it’s just me,” she tells her therapist. Acknowledging that the work has made her panic, she explains, “It’s an unfortunate, manageable symptom of doing something important.”

Across the genre, some characters nurse guilt over harm they’ve caused in the past and take the job to atone: Margot, in Faces of Death, says that she wanted to “do my part, and try to clean up the internet,” in the hopes that “something good can come out of” her sister’s getting hit by a train while they were filming a dance video together. In Sky Daddy, a colleague of Linda’s says she took the job out of a desire to “punish” herself for bullying a girl who later died. The female protagonists of these works generally come into their jobs with trauma and are then subjected to more trauma on the job, which in turn pushes them closer to a breaking point.It is true that many of these real-life workers struggle greatly; in some cases, they reportedly have developed PTSD. But fictional portrayals of these low-paid workers as ill-adjusted victims advance a niche version of the trauma plot — one that flattens its characters and circumscribes their worlds to the point of claustrophobia. 

Perhaps because their jobs, and their identities, are so geared toward the screen, several fictional content moderators are given the chance to star in controversial videos — and even be moderated — themselves. Jane is in therapy because a video of her having a breakdown at the office has reached a wide audience: “I was meme,” she says. “Today, on my way here, a man asked if he could record me saying happy birthday to a friend.” Margot is recognized at a convenience store by a young woman who asks: “Are you the girl from that train video?” In American Sweatshop, a colleague alerts Daisy that he has been tasked with reviewing a video of her beating up a bartender, captioned “Girl Boss DESTROYING that guy into PULP bloody mess (security cam).” A coworker of Linda’s alerts her, likewise, that he “saw a crazy video today” of her jerking off her boss “with unseemly gusto” on a plane. Kayleigh’s ex-girlfriend, in We Had to Remove This Post, shows her a video of their tryst in a supply closet and demands: “You see this at work — what do you do: Leave it up or take it down?” Kayleigh considers: “Sexual content,” but with “No female areola, no genitalia, so leave it up.” 

As anyone who writes about tech has had to learn, what happens on a screen, even in the most absorbing possible configuration, is generally boring to read about or watch. In a tacit acknowledgment of this fact, several writers of content moderator fiction have their protagonists try to take matters into their own hands IRL. The climax of Job sees Jane — who has made a habit of using her advanced internet sleuthing skills track down the “freaks” who are “living among us” — accuse her therapist of being an abuser from videos she’s seen, in a plot twist so galling that, as Jesse Green wrote in his New York Times review, it feels “even more manipulative than other therapy-based psychological thrillers.” We Had to Remove This Post ends with the protagonist sneaking into the family home of a girl who had uploaded self-harm videos. In the final scene of Faces of Death, Margot streams herself slicing up the viral sicko she’s been tracking. Girlie, in Moderation, is shaken when she thinks she has seen a man from disturbing posts walking by her outside a casino. These shifts in the action generate excitement, but they also demonstrate a fundamental misunderstanding of how these jobs, and the internet, actually function: they depend on the idea that the internet is small when, in fact, it is vast and unwieldy. As cathartic as a confrontation may be, it is unlikely a moderator would run into evil people out and about, or even live in the same region as them. Content moderation fiction falters when it refuses to accept that part of what makes this work so brutal and dehumanizing is that it is almost entirely random. 

The idea that a small group of unlucky people could absorb or repel the darkest elements of the web, leaving social media sites pleasant for the rest of us, was once mainstream. Now, the conversation has shifted from the risks of bad posts on the platforms to the dangers of the platforms themselves. Judicial and legislative cures have grown sweeping, seeking as they do to limit young people’s use of the platforms altogether. A judge in New Mexico, for example, ordered in August that Instagram and Facebook could no longer send push notifications to underage users between 10 p.m. and 7 a.m., and that it must limit those teens’ monthly usage. At issue here, and in the thousands of lawsuits Meta faces for causing user harm, is not just the noxious content someone might see on a feed, but access to the feed itself, including features like infinite scroll. The entire model, in this new formulation, is toxic. That a human barricade between users and the feed could protect us from the dangers of social media now seems quaint — as does the notion that the platforms, always beholden primarily to advertisers, would act in users’ best interests, insulated from political pressures. The optimistic mindset of the 2010s has curdled; no amount of moderation would undo a calcifying sense that the platforms are irredeemable. 

That doesn’t mean that content moderators no longer exist. Though social media companies continue to shrink their moderation teams, opting to slot in A.I. wherever possible, humans are still brought in for the most complicated, often most brutal, tasks. In a March blog post announcing new A.I. features, Meta said that it would “reduce our reliance on third-party vendors for content enforcement” but that “people will continue to play a key role in how we make the highest risk and most critical decisions.” In August, a TikTok moderator stepped in to review a livestream of Perez Hilton appearing to harm himself after the clip had been flagged by A.I. 

And A.I. has already presented opportunities, if they can be called that, for contractors to wade into the nastiest, most disturbing content. A new class of workers has been brought in to flag and filter out distressing material from appearing in A.I. chats. Workers in Kenya, for example, were paid two or three dollars an hour to comb through racist vitriol generated by new A.I. tools. A woman hired to be a “generalist rater” to train Google’s Gemini told The Guardian that she soon found that her shifts exclusively involved flagging lewd and violent content for removal. Mophat Okinyi, a moderator in Nairobi who screened disturbing sexual content on ChatGPT for less than four dollars an hour — and went on to cofound the African Content Moderators Union — told the reporter Karen Hao in 2023 that “I’m very proud that I participated in that project now to help keep ChatGPT safe. But now the question I always ask myself: ‘Was my input worth what I received in return?’” 

Fiction about content moderators aimed to capture and critique the moment, but that moment has evolved more quickly than any of its creators could have anticipated. Even a couple years make a difference: Faces of Death, the most recent release of this batch, was supposed to premiere in 2024, but after various delays came out this spring. Its star Barbie Ferreira told an interviewer that the movie “has only become more relevant through these years. It’s just been so interesting to see the world evolve into even more of this algorithmic violence hellscape that we live in.” Her spin is admirable, and she’s correct that the world only has become more hellish. But a more apt word there might be dated. The project, and others like it, characterize a world we no longer live in, as the relationship between the platforms, the workers, and the public has been scrambled. 

Many writers have, in their careers, seen projects that began as dystopian fiction end up more like realism. Vauhini Vara, for example, has said of her 2022 tech novel The Immortal King Rao that “reality moved closer to what I imagined as a sort of speculative future in my book.” Writers of content moderator fiction have swiftly seen their projects undergo a different turn: from contemporary realism to historical fiction, as aggressive moderation has withered into permissive policies, and A.I. has overtaken workers’ tasks. The playwright Max Wolf Friedlich, in the author’s notes for Job, makes an acknowledgment that might — given both subject matter and posture toward the industry — be appended to all of these works: “This is a period piece.”