X Revenue Sharing Rewards AI-Generated Fake Melodramas
The Transparent Lie
X publishes its rules openly. The Creative Revenue Sharing program states a simple ambition: to reward “authentic, high-quality content” that boosts engagement on the platform. The thresholds are documented for anyone to read. A creator needs Premium status, at least 500 verified followers, and a minimum of 5 million impressions over three months. That is transparency in its most literal form. Explainability is a different creature altogether.
Explanation requires tracing what actually happens once those rules meet human ingenuity. Following the money, the prompts, and the psychology is essential. It requires sitting with the uncomfortable realization that a platform can be fully transparent about its mechanisms while the outcomes remain almost perfectly opaque to the people scrolling through them. The original investigation into this phenomenon, published by WIRED, traced one such story back to its source and found a machine built on deception. [1]
Consider the story that opens that investigation. A woman sits in federal court, certain she will lose her freedom. Her husband has spent six months convincing everyone, including their daughter, that she stole millions from the company she built. Then the courtroom doors open. Her nine-year-old son walks in carrying an old blue backpack, looks at the judge, and announces that he knows who framed his mother. The culprit is sitting in the room.
The narrative is engineered for maximum emotional impact. It has a villain, an innocent victim, a child savior, and a courtroom showdown. It also has no basis in reality. The first-person narrator is not a mother on trial. The author is one of countless beneficiaries of X’s Creative Revenue Sharing program: a self-described 21-year-old stock trader in Nigeria who uses AI to generate stories and receives payments for doing so. The fiction is presented as memoir, and the platform distributes it to millions of feeds, including mine. The posts are manipulative and dumb and anything but authentic.
This is the heart of the deception. The lie is not merely that individual stories are fabricated. The deeper fabrication is the impression that these are authentic voices sharing lived experience. The program’s transparency about its existence and rules masks the absence of any meaningful verification about what is real and what is generated.
The Incentive Machine
The economics of this deception are brutally simple. X pays creators every two weeks as part of the Creative Revenue Sharing program. The Nigerian trader who posts under the handle BIGBEN says his paychecks usually land between $500 and $700 per cycle. Some creators report earnings of thousands of dollars per payment period. For a young person in Lagos, that is not pocket money. That is a livelihood.
The scale of the operation becomes clear when you look at the numbers attached to these posts. A single story thread posted by BIGBEN, narrated by a second wife abused by her in-laws, notched 1.5 million views in two days. [1] A courtroom melodrama routinely garners hundreds of thousands of views, hundreds of likes, and dozens of reposts. Some of that traffic is certainly bot-driven. Much of it is real, human, and hungry.
The structure of these stories follows a pattern as formalized as a sonnet. Rob Cover, a professor at Australia’s RMIT Digital Ethnography Research Centre who studies AI-generated content, breaks it down with clinical precision. [2] The first movement is a rage-baiting setup in which an innocent person is wrongfully fired, evicted, scammed, or stripped of something precious. [2] The second movement is a turnaround: the underdog is redeemed while the adversary is defeated, humiliated, and shamed. [2] The formula is consistent because it works.
Content monetization programs have been gamed since the earliest days of web advertising. What changed with AI is the cost of production. A human writer needs time, energy, and some relationship to experience to produce a viral story. A language model needs a few seconds and a prompt. The barrier to entry collapsed, and the pool flooded.
The data pool that feeds this machine is the platform’s own attention economy. Every click, every idle engagement, every moment of curiosity feeds the algorithm information about what keeps people watching. The quality of that pool determines the quality of what gets produced. When the algorithm rewards melodrama, the pool becomes saturated with it. The incentive structure converts the platform into a factory for fabricated suffering and scripted revenge.
The deception operates on a second level here, aimed at the platform itself. Creators are not merely deceiving readers. They are deceiving X’s economic engine. The platform moved to cut payments to aggregators who primarily posted news from other sources in April 2025. It threatened sanctions against those who overused the BREAKING tag to draw eyeballs. In May 2025, X sued eight Vietnamese creators and 25 unidentified John Does, alleging they fraudulently drew money from the program by posting phony AI-generated content and boosting engagement with bot-driven comments. On July 16, 2025, X revealed it had removed millions of posts that purloined content already published by others. The platform is fighting back against the contamination of its data pool, and yet the melodramas keep coming.
The people running these operations have optimized for precisely the wrong incentives, which is to say the precise incentives the platform built. Jenna Russell, a PhD candidate at the University of Maryland studying AI writing, describes the process as blindingly simple. [3] Spammers, she says, have simply optimized for the least effort and the most engagement. [3] That means using easy, well-known tropes and prompting AI to write something that fits the bill. The result is a genre that did not exist five years ago and now dominates the feed.
Stories in this genre recycle a handful of reliable scenarios. Narrators often wield sealed documents that turn the tables on tormentors in divorce disputes or battles over a relative’s estate. Sometimes the drama is smaller: an airline passenger who fiercely retains a seat when rude moochers claim it, sending the adversary on a walk of shame down the aisle. The variations are endless, the underlying structure identical.
The Hollow Payoff
A consequence follows necessarily from what the incentives establish. When the algorithm rewards engagement without regard to truth, the pool fills with what performs. When performance is measured in views and comments, fabrication becomes the rational strategy. Its own metrics become the engine of its own degradation.
Writers who actually produce original content are the first casualties. A platform flooded with AI slop makes it harder for real voices to be heard. The attention economy has a finite budget. Every minute spent reading a fabricated courtroom saga is a minute not spent on something real. The deception does not merely fool people in the moment; it displaces the authentic material that might have earned their attention honestly.
Readers themselves are also deceived. These stories are presented as true first-person narratives. Some creators describe themselves in their profiles as storytellers, a form of honest labeling that many others omit. The reader who believes she is witnessing a genuine account of injustice and redemption is being manipulated by a machine optimized to exploit her psychological vulnerabilities. Reading them delivers a jolt of dopamine as justice prevails, and the AI knows exactly how to keep that jolt coming.
The company behind X, now folded into xAI and then into SpaceX, declined to answer questions about how the platform regards these AI-generated story posts. That silence is itself informative. A platform that pays for engagement cannot easily acknowledge that its payment system produces something hollow at its core. Efforts to confirm how X views the slop that has infected so many feeds went unanswered.
Reaching BIGBEN required persistence; several other creators ignored requests to explain their methods. He contended that some of his stories came from experiences people shared with him, but he freely admitted to using AI for the actual writing. When a new story is needed, he gives Grok or ChatGPT a prompt and lets the model do the rest. His initial response revealed a fundamental misunderstanding of journalism. He proposed a partnership, uploading his stories to a publication in exchange for a cut of the revenue. When told the scheme would not work, he turned cautious. “Before I give out my leverage, what’s in it for me?” he asked. [1] That transactional logic defines the entire economy: every exchange is a deal, every interaction a potential revenue stream.
The popularity of these tales reveals something genuine. Hunger for fairness, lust for retribution, satisfaction in seeing villains shamed and underdogs vindicated, all of these are real human needs crying out for satisfaction. The AI does not invent those needs. It identifies them and feeds them. It produces content our brains cannot resist, at scale, without fatigue, without conscience, and without any stake in whether the justice it delivers exists anywhere outside the words it generates.
The courtroom story ends neatly. A nine-year-old boy planted a digital recorder in his father’s office. The recording exonerates the mother and puts the husband behind bars. The denouement is satisfying. The reading experience is clean and complete, entirely hollow once you know none of it happened.
That hollow feeling may become the signature emotional experience of the AI era. Deception operates most completely when it conditions us to accept fabricated emotional experiences as substitutes for real ones. Gradually, it trains us to consume justice, suffering, and redemption as content, detached from any actual human experience.
Attempts to clean up the mess demonstrate that the problem is structural, not incidental. Lawsuits, policy changes, and post removals address symptoms while the incentive machine keeps manufacturing the disease. The Creative Revenue Sharing program still pays for engagement. The algorithm still amplifies what keeps people watching. The data pool still rewards the formula.
Each improvement in generation quality makes the fabricated stories harder to detect. Each refinement of the formula makes the emotional manipulation more precise. The machine learns from every click, every share, every indignant comment, and turns those reactions into better bait.
What remains to be seen is whether the platform can escape this trap. Rules have changed since the program began. Payouts have been adjusted. Lawsuits have been filed. Removals have been executed. All of these measures fight the last war. The next generation of AI-generated content will be harder to distinguish from authentic human expression, and the incentive to produce it will remain as long as engagement pays.
Readers of the courtroom saga did not know they were being deceived. The platform that paid for it may not fully grasp the depth of the problem. The creator who wrote it has no reason to care. He is 21 years old, producing stories by prompt, and cashing checks every two weeks. The machine works exactly as designed. The hollow feeling that follows every revelation of fabrication is the cost of admission. We want justice. The AI knows it. The platform profits from it. The truth, as always, arrives last.
Sources
1. WIRED
2. RMIT Digital Ethnography Research Centre
4. SpaceX