The Same Room, Rebuilt

Part 01

Part One: The room gets built (1994 to 2005)

The pattern predates video. It predates broadband. It begins the moment strangers could type at each other in real time.

America Online's chat rooms in the mid-1990s were the first mass-market space where adults and children shared unmoderated real-time contact, and the consequences arrived almost immediately. In 1995 the FBI launched Operation Innocent Images, its first major national initiative against online child exploitation, after investigations found offenders using AOL rooms and private messages to groom children and trade abusive imagery. The playbook agents documented then is, beat for beat, the one platforms are still fighting today. An adult enters a youth space. He offers attention, flattery, the feeling of being seen. He moves the conversation somewhere private. He escalates his requests slowly, then uses shame and secrecy to hold the child in place. The corporate response established the cycle's second law: platforms act when the liability exceeds the engagement, and not a moment sooner.

In September 2003, Microsoft shut down unmoderated MSN chat rooms across 28 countries, citing the danger of predators approaching children, one of the first times a tech giant amputated a popular feature explicitly on safety grounds. Yahoo followed in 2005, killing user-created rooms after reporters found rooms whose names openly advertised sexual interest in minors, and after major advertisers pulled their money. The rooms had not suddenly become dangerous. They had suddenly become expensive. The mid-2000s gave the public a comforting villain. NBC's To Catch a Predator turned chatroom grooming into prime-time entertainment and fixed the image of the internet predator as a sweaty man walking into a sting house. MySpace, under pressure from state attorneys general, purged tens of thousands of registered sex offenders from its rolls. The lesson America absorbed was that the problem was a finite list of bad men who could be caught and shamed.

The lesson the data actually taught, that the danger was architectural, built into any system that makes minors visible to anonymous adults, went unlearned. It stayed unlearned for twenty years.

Part 02

Part Two: The webcam turns on (2005 to 2013)

Stickam is the most important website most people have never heard of. Launched in February 2005, months before YouTube found its footing, it let anyone broadcast a live webcam feed to strangers with a chat box beside the video. By 2012 it claimed over nine million registered users, and its culture was unmistakable: teenagers, broadcasting from their bedrooms, performing for an audience that talked back. Stickam invented, mostly by accident, the entire grammar of modern social video. The parasocial bedroom broadcast. The viewer-request dynamic, where the chat tells the performer what to do next. The follower gate. Even a random-pairing shuffle feature that beat Chatroulette to its own gimmick. And within two years it had demonstrated exactly what that grammar attracts. Child-safety advocates were flagging the platform by 2007.

In 2009 alone, American authorities made multiple arrests tied to the site, including a man who posed as a teenage boy to deceive underage girls into sexual acts on camera, amassing over a hundred videos, and another who coerced a fourteen-year-old into on-camera acts before arranging to meet her. Reporters also traced connections between Stickam's parent

company and adult webcam businesses, an early glimpse of a boundary that would spend the next two decades dissolving: the line between the teen social platform and the cam site. The Stickam era also created the thing that makes livestream exploitation permanent. Live video is ephemeral, so a subculture formed around recording it. Viewers captured broadcasts, especially the moments when young streamers were dared, coaxed, or tricked into exposing themselves, and archived the footage for trading elsewhere. Law enforcement now calls this "self-generated" child sexual abuse material, a clinical term for content a child technically produced herself, under manipulation, which then circulates forever beyond any platform's reach. The UK's Internet Watch Foundation would eventually report that self-generated imagery, much of it showing girls aged eleven to thirteen in their own bedrooms, had grown to make up the majority of the abusive material it removes from the internet.

Every one of those images is the long shadow of a chat box that said do it, and a child who wanted the room to like her. Stickam died in January 2013 with one day's notice, officially of obsolete technology. Its audience, on both sides of the camera, simply moved. Where they moved was Omegle. Launched in 2009 by an eighteen-year-old, Omegle stripped the model to its rawest form: press a button, face a stranger on video, no account, no name, no age check. For years its splash screen carried an astonishing sentence, warning users that predators were known to use the site. It read less like protection than like a liability disclaimer, and the site eventually deleted it. The design was the problem, and everyone knew it. A CBC reporter who used the platform in 2022 found that most strangers she was paired with appeared to be men who were naked or deliberately off camera.

That same year, Omegle, a site run by a skeleton crew, filed more than 600,000 reports of apparent child sexual exploitation to the National Center for Missing and Exploited Children, more than TikTok, Snapchat, or Discord. The lawsuit that finally killed it reads like a schematic of the entire cycle. In 2014, an eleven- year-old girl, identified in court as A.M., was randomly paired on Omegle with a Canadian man in his late thirties. According to her complaint, he obtained her contact information and spent the next three years exploiting her, forcing her to produce abusive material for him and his friends, and eventually directing her to go back onto Omegle to recruit other girls for him. The victim became the recruiting tool. When she sued, Omegle raised the defense every platform raises, Section 230, the American law that shields sites from liability for what users post. The judge refused it.

The harm, he found, flowed from the design of the product itself, from the decision to randomly pair unverified children with unverified adults, and design is not speech. Omegle settled in November 2023 and shut down within days. Its founder's farewell letter conceded the site had been used for "unspeakably heinous crimes" and, as a condition of the settlement, thanked A.M. by name. Even in death, Omegle proved the cycle's central theorem. Child-protection monitors watching dark-web forums recorded users mourning the shutdown and openly discussing where to find children next. A lawyer for another child victim put it plainly to Forbes: closing the site does not change the behavior. The demand migrates. It always migrates.

Part 03

Part Three: The money arrives (2013 to 2021)

The smartphone changed the physics. Stickam's teenagers performed for attention. The mobile generation performed for attention that converted into currency, and that conversion is the hinge on which everything since has turned. YouNow inherited the teen broadcast audience in 2011 and pioneered gifting: virtual items bought with real money, sent to broadcasters, partially cashable. Live.me, launched in 2016 and promoted heavily through Musical.ly, the app that would become TikTok, ran the same economy and became notorious when investigations found adults using gifts to solicit sexual displays from visibly underage streamers. Kik became the era's connective tissue, an anonymous messenger so saturated with grooming that police forces on three continents described it as a fixture of their child-exploitation caseloads.

Periscope, Twitter's shiny 2015 acquisition, showed what happens when a major corporation inherits the pattern and stops paying attention: by late 2017, Gizmodo found dozens of accounts openly soliciting sexualized imagery of minors, some using explicit images of children as avatars, many with thousands of followers and zero public broadcasts, because the activity had moved into Periscope's private groups. Twitter announced it had banned 36,000 accounts that year for inappropriate engagement with minors, a statistic offered as reassurance that functioned as a confession of scale. Periscope died in 2021, officially of cost-cutting. And then there was Peeks Social, a small Canadian app that deserves a footnote for its honesty. Built from the remains of a failed video service, Peeks made tipping and pay-per- view the entire product from day one.

It drifted immediately toward sexualized content, because sexualized content is what live tipping reliably buys, fought with app stores and payment processors, and collapsed. Peeks was every other platform's life cycle run at fast- forward, with the denial removed. By 2021 the lifecycle could be written as a formula, and it is worth writing out, because every platform in this story has run it. Stage one: a niche platform launches with a novel social mechanic and no meaningful safety infrastructure. Stage two: young people colonize it, because young people colonize everything new, and because it offers what adolescence craves, which is visibility. Stage three: adults with money arrive, drawn precisely by the youth of the room, and begin shaping behavior through attention, dares, and payment.

Stage four: the platform formalizes the money into gifts, coins, and tips, takes a percentage of every transaction, and becomes financially dependent on the dynamic destroying its users. Stage five: performers escalate, because escalation is what pays; journalists or litigators force a reckoning; the platform dies or pivots; the audience disperses to stage one of the next platform. Omegle, Stickam, Live.me, Peeks, and Periscope all ran the full course. The question of this decade is why TikTok hasn't. The answer is uncomfortable: it didn't break the cycle. It franchised it.

Part 04

Part Four: The strip club goes corporate

In April 2022, Forbes reporter Alexandra S. Levine published an investigation into TikTok Live that should have been a extinction-level event for the feature. She documented, across hundreds of streams, adult viewers using the comments to direct teenage girls broadcasting from bedrooms and bathrooms through escalating requests, and rewarding compliance with virtual gifts redeemable for cash, or with payments routed to the Cash App and Venmo handles the girls listed in their bios. The viewers had developed a whole vocabulary of innocuous-sounding prompts engineered to get girls to display their bodies to the camera, requests that sound like games and function like commands. A Harvard Law School dean quoted in the piece described the feature as the digital equivalent of a strip club staffed by fifteen-year-olds.

A Homeland Security agent explained the economics of the other side of the screen: a ten-dollar gift is a spectacular investment for a predator, because it gets a child doing something she would not otherwise do, and once she has done it, the leverage exists. That is the sextortion pipeline, and live gifting is one of its widest mouths. What elevates TikTok from scandal to indictment is what the company did next, which we only know because state attorneys general sued and their filings were imperfectly redacted. After the Forbes story, TikTok ran an internal investigation code-named Project Meramec. It confirmed everything. In one single month of 2022, roughly 112,000 underage users hosted livestreams in violation of the platform's own age rules. Viewers sent about one million gifts to minors engaged in what the company blandly called transactional behavior.

The recommendation algorithm was found to be actively boosting the sexualized streams, because they generated the most engagement, and TikTok collected its commission on every gift. A second internal probe, Project Jupiter, found the same gifting rails being used to launder drug money. One internal message, quoted in the litigation, compresses thirty years of this history into a single sentence: the content generating the highest engagement was not the content the company wanted on its platform. More than a dozen states sued in 2024 and 2025. Utah's complaint called TikTok Live a virtual strip club. Minnesota's alleged the company was operating an unlicensed money- transmission system that profited from the sexual and financial exploitation of children. TikTok denies wrongdoing and points to its policies. But notice what did not happen. TikTok Live was not shut down. Nobody's farewell letter thanked a victim by name.

The feature that an internal investigation confirmed was monetizing minors is still running, at global scale, right now, because TikTok is not a niche site that can be sued into oblivion. It is one of the largest media products in human history, and the gifting economy is load- bearing. This is the modern era's real innovation. The behavior that used to kill platforms has been licensed. Age gates, launched under regulatory pressure like the UK's Online Safety Act, whose mandatory age checks took force in July 2025, now segregate the officially sexual economy from the officially all-ages one. Paid private rooms and subscriber-only broadcasts move the highest-risk interaction out of public view, which protects the platform's image far more reliably than it protects anyone inside the room; Periscope's

private groups demonstrated a decade ago that privacy features are where the worst activity concentrates. Screen-capture deterrents and watermarks attempt to slow the secondary market in recorded streams, and the continued existence of entire trading communities built on captured live content tells you how that contest is going, without anyone needing to explain the methods. Each of these mechanisms genuinely reduces some harm at the margin. Each also functions as liability engineering: the paperwork that lets the gifting economy continue.

Part 05

Part Five: Novinha, and the girls the algorithm exports

Everything described so far gets worse when it crosses into the Global South, because every input of the cycle intensifies. The demand side gets richer relative to the supply side. The safety infrastructure gets thinner. And the cultural machinery for sexualizing girls is older than the internet and already set to music. Brazil is the clearest case study on earth, and the vocabulary tells the story before the statistics do. Novinha is an ordinary Portuguese diminutive. It means, literally, young girl. It is also one of the most common words in Brazilian funk, a genre whose lyrics have spent two decades attaching it to desire, and which soundtracks the dance videos that Brazilian children post by the million. The word has become a genre, a search term, and a market signal all at once.

When a girl in João Pessoa or Belém posts a dancinha, a little TikTok dance in her school uniform, the word waiting in the caption space and the comment section is one the culture has already sexualized for her. She did not choose the frame. The frame was there first. The live-request culture completes the machine. In Brazilian streams and comment sections, one of the recurring demands made of girls is the 360: turn around, all the way, slowly, so the camera can see everything. It presents as a compliment, an appraisal, a game, exactly like the "outfit check" prompts Forbes documented on English-language TikTok. It is an inspection ritual, the showroom turn, and girls learn to perform it before they are old enough to understand what is being appraised. This is what adultização means, the word Felca put on forty million screens: not that children are exposed to adulthood, but that they are processed into it early, formatted as content for a market that was waiting.

The Hytalo Santos case showed how far the format scales. Santos, an influencer with eighteen million followers, built what prosecutors described as a reality-show universe around a rotating cast of minors, filming them in relationships, in sexualized dances, in adult situations, and monetizing all of it. One of the girls had been in his content since she was twelve. He had been under investigation since 2024, after neighbors reported parties involving alcohol and children, but it took Felca's video to make the case nationally unignorable. Santos was arrested nine days after the video posted. He was later convicted, and he and his partner also face labor-court proceedings for human trafficking and conditions analogous to slavery. His defense argued, and this is the detail worth sitting with, that everything was filmed with parental authorization. In the adultização economy, the parents are frequently inside the production. Felca's video also documented a mother

selling explicit material of her own teenage daughter through Telegram because the followers kept asking for more. The numbers around the scandal are staggering. SaferNet Brasil logged nearly 77,000 reports of online rights violations in the first seven months of 2025, and 64 percent of them involved child sexual abuse and exploitation. After Felca's video, reports of child sexual abuse material more than doubled, which is what happens when a country suddenly learns what to look for. Congress, shamed into motion, passed the ECA Digital, a digital child- protection statute that had been stalled since 2022. Article 23 of the law bans platforms from monetizing or algorithmically boosting content that portrays children in eroticized or sexually suggestive ways. It is, on paper, one of the strongest laws of its kind in the world. And here is what happened next, because the cycle does not read legislation.

In the first week after the law took force in March 2026, Brazilian journalists checking Meta's own advertising library found boosted, paid, platform-approved ads promising leaked content of novinhas and students. Predatory accounts continued trawling years-old videos of young girls, leaving comments in Portuguese, Spanish, and English, because the demand for Brazilian girls is international, and a dance video posted in December is still a point of contact in March. The law changed the liability. It has not yet changed the market. Brazil is the loudest example, not the only one. The same dynamics run through the gifting apps of Southeast Asia, where investigations have repeatedly found broadcasters in poorer countries performing for viewers in richer ones.

At the extreme end sits the on-demand livestreamed abuse economy that law enforcement has been dismantling in the Philippines for over a decade, where joint operations found networks selling live abuse of children to Western customers, with tens of thousands of dollars flowing from the customer network for acts priced, individually, at less than a night out. The distance between a gift-baited 360 on a public stream and a purchased private show is a spectrum, not a wall, and the entire apparatus of global inequality pushes girls down it.

Part 06

Part Six: The countdown, or why the pipeline feels inevitable

Now follow one girl all the way through, because the pipeline only makes sense as a biography. She is thirteen when she posts her first dance video. The algorithm, which has learned what Felca proved it learned, gives it more reach than anything she has ever made, because the audience for thirteen-year-olds dancing is large, motivated, and highly engaged. The comments teach her the rules: this angle, that movement, this kind of top. At fourteen she goes live, and the gifts start, small ones, roses and ice cream cones, attached to small requests. She learns, the way every broadcaster since Stickam has learned, that compliance is revenue and hesitation is churn. At fifteen, some of her clips have been captured and reposted to accounts and threads she will never see, where men annotate her age. At sixteen she has a following that is openly counting.

The eighteenth-birthday countdown is a real, documented, public ritual now: audiences tracking a girl's age like a product release

date, waiting for the day her sexuality becomes legally sellable, and telling her so in the comments. Her adulthood arrives pre-sold. So on her eighteenth birthday, or near enough, she opens the subscription page, and everyone acts surprised. The infrastructure receiving her is enormous and mainstream. OnlyFans processed 6.6 billion dollars in payments in 2023. Its marketing layer is X, formerly Twitter, the only major Western platform that openly permits pornography, a position it formalized in written policy in 2024. The mechanics are frictionless and familiar because she has been rehearsing them for five years: post free teaser content on X and TikTok, put the Linktree in the bio, funnel the audience from the all-ages platform to the paid one. The link in bio is the entire architecture of the modern pipeline, a single hyperlink that connects the platform where she built an audience as a minor to the platform where that audience pays.

A 2020 BBC investigation found that on a single day, roughly a third of the Twitter profiles worldwide advertising nudes for sale appeared to belong to underage people, many of them funneling to OnlyFans. The BBC reported in 2021 that OnlyFans was failing to keep underage creators off the platform. Reuters reported in 2024 that predators were still exploiting children there. The front door has ID checks. The building has always had other doors. And here is why, for some girls, the pipeline is not a choice in any meaningful sense but a gradient, which is what people mean when they say it feels inevitable. Consider what she has by eighteen. She has years of conditioning that attention from adult men is income. She has an audience that was assembled around her body before she could legally consent to selling it, an asset she did not build on purpose but which is the only asset the market has let her build.

She frequently has captured content already circulating, which collapses the psychological barrier; the thing has already been taken, so selling it at least returns some of the value. She often has real economic need, and in Brazil or the Philippines, catastrophic economic need, in an economy where her follower count is worth more than any credential she could earn. And she has a culture, which we will get to, telling her that this is empowerment. What she does not have is good odds. The average OnlyFans creator earns around 1,300 dollars a year. The distribution is brutally top-heavy, which means the platform's economics run on volume at the bottom: millions of creators earning almost nothing, each incentivized to escalate, discount, and do the thing she said she'd never do, because escalation is the only lever the long tail has.

Guardian and BBC investigations in 2026 documented the agency layer that has grown on top of this desperation, male-run management firms taking half or more of young women's earnings, ghost-writing their intimate chats through employed "chatters" and bots, trading the women between agencies like accounts, and in the worst cases shading into coercion outright. The empowerment platform reinvented the pimp, gave him a dashboard, and called him management. That is the pipeline entire. A girl monetized at thirteen by a platform's algorithm, harvested at fifteen by a capture economy, counted down at seventeen by her own audience, onboarded at eighteen by a subscription giant, and skimmed at nineteen by an agency. At

every stage, someone other than her took a percentage. The internet did not invent any single step. What it did was connect them, lubricate them, and put a progress bar on them.

Part 07

Part Seven: The death of shame

There is one more component, and it is the one that makes the machine culturally self- sustaining: the story we all agreed to tell about it. Somewhere in the last fifteen years, roughly between the fall of Stickam and the rise of OnlyFans, the public conversation about all of this was captured by a vocabulary of liberation. Selling access to your body online became sex work, and sex work became work, and work became empowerment, and empowerment became something it is rude, prudish, and vaguely right-wing to question. Platform payment processors are censorship. Age verification is surveillance. Criticism of the industry is stigma, and stigma is violence. Every guardrail, in this vocabulary, is an attack on someone's freedom of expression, and every twenty-year-old with a subscription page is an entrepreneur. Some of that vocabulary answered real cruelty.

The women in this economy did not invent it and are not the villains of it; shame, historically, has been a weapon aimed at them while the demand side walked away clean. Adult women genuinely do choose this work, some profitably and freely, and a discourse that treated them as fallen was worth burying. But notice what the death of shame conveniently killed alongside it. Shame was doing a job. It was crude, it was cruel, and it was misallocated, landing on women instead of on the men buying and the platforms skimming. But it was the only price the demand side ever paid. The man who follows a fifteen-year-old's countdown, the agency taking sixty percent, the platform boosting the dancinha, the advertiser whose ads run against it: the collapse of stigma cost them nothing and freed them completely. We did not redistribute the shame to the people who deserved it. We abolished it, and the beneficiaries were the buyers. The freedom-of-speech framing does similar work.

It relocates a market into the category of expression, where liberal societies are correctly reluctant to regulate, and it dares regulators to look like censors. Omegle's founder, in his farewell letter, compared efforts to hold his platform accountable to restricting women's clothing to prevent rape, casting a site that had generated six hundred thousand exploitation reports in a year as a martyr for civil liberties. That is the move, made with unusual bluntness. Every age gate becomes a privacy debate. Every design regulation becomes a speech debate. The A.M. ruling mattered precisely because it refused the frame: pairing children with anonymous adults is not speech, it is a product decision, and products can be defective. And the consequences, in this new dispensation, are supposed to have died too.

That is the deepest false promise the culture makes to the girl at the top of the pipeline: that it is just content, that she can stop whenever she wants, that the internet forgives. Every part of that is a lie, and the machine itself proves it. The capture economy exists because content is forever. The sextortion statistics exist because a single image is leverage; the National Center for Missing and Exploited Children received 1.4 million reports of online enticement

in 2025, up from 44,000 in 2021, and it knows of at least three dozen American teenagers, mostly boys, who killed themselves after being extorted with images they were manipulated into sending. Generative AI has now removed even the need for the image to be real; reports involving AI-generated material went from under five thousand in 2023 to over 1.5 million in 2025, meaning any child with a public school photo is now, potentially, raw material. The consequences did not die. They were privatized, moved off the ledgers of the platforms and the buyers, and billed entirely to the person on camera. This is what people are gesturing at, clumsily, when they say this is damaging the fabric of society, and the phrase deserves to be rescued from the culture war it usually lives in. The damage is specific and observable.

A generation of girls trained by reward schedules to price their own visibility, and a generation of boys raised inside the buyer's side of the interface, learning that women come with a tip menu. Intimacy reframed as subscription retention, complete with outsourced chat girlfriends who are actually men in an office, or software. Parents converted into managers of their children's sexualized output, because the household needed the money and the platform paid. Communities, in mainstream corners of the internet as well as hidden ones, organized entirely around cataloguing girls as they age toward legality. None of this is a moral panic about young people having sexuality. It is a market description. The panic would be not panicking.

Part 08

Part Eight: What thirty years actually taught us

It would be false to say nothing has improved. The A.M. ruling cracked Section 230's shield by aiming at product design instead of user content, and it killed a platform. Brazil's ECA Digital, the UK's Online Safety Act, the American REPORT Act and TAKE IT DOWN Act, and the state attorney-general lawsuits amount to the first sustained regulatory pressure this industry has ever faced. Detection tooling, hash matching, and mandatory reporting pipelines are real infrastructure that the Stickam era simply lacked. Felca's video showed something genuinely new, too: a creator using the platforms' own reach to indict the platforms, and winning, at least for a news cycle. But the lifecycle itself has never been broken, because its inputs have never changed. Young people's hunger for visibility. Adult demand for access to them. Anonymity, or its modern equivalent, plausible deniability at scale. And a monetization layer that pays the platform a commission on the collision.

Every previous solution removed a venue and left the market intact. The Canadian Centre for Child Protection said it plainly after Omegle fell: without structural regulation, we are simply waiting for the next site to backfill the last one. Brazil said it again, involuntarily, when the novinha ads reappeared in Meta's ad library the same month its landmark law took effect. The honest conclusion of thirty years of evidence is that this was never a moderation problem, and it was never a bad-men problem, though the bad men are real and the moderators are outgunned. It is a business-model problem. Any system that pairs minors' visibility with adults' money and takes a percentage will produce this outcome, reliably, measurably, in every language and hemisphere, regardless of the sincerity of its trust-and-

safety team. The cycle ends when that pairing becomes legally and financially impossible, when the commission itself is the crime. Until then, the room keeps getting rebuilt. New logo, new currency, new country, same chat box, same bedroom, same audience. And somewhere right now, on a platform you have not heard of yet, the comments under a thirteen-year-old's dance video are already teaching her the rules.