X Is Paying Strangers to Flood Your Feed with Fake Heartbreak

X built a marketplace that pays creators for engagement, set a 5-million-impression floor for entry, and asked the global internet to figure out what kind of content pays. The answer turned out to be AI-generated first-person heartbreak falsely presented as memoir, and three rounds of X enforcement since May 2025 have all targeted something else.

The program at the center is X's Creative Revenue Sharing, a per-engagement payout system that requires a Premium subscription, at least 500 verified followers, and 5 million impressions over the trailing three months. The thresholds are not unusual for creator monetization. X chose to monetize a recommender trained to maximize the same engagement it pays out on. Every tearjerker a user clicks trains the algorithm to surface more tearjerkers, and every surfaced tearjerker pays a creator to make another.

A self-identified 21-year-old stock trader in Nigeria has used AI to mint first-person confession threads that WIRED traces to a single handle, fwBen_. One of his posts drew 1.5 million views in two days, according to WIRED's reporting on the AI slop genre: mass-produced short fiction falsely presented as memoir. The genre's signature is its plausibility. A child addresses a federal judge. A stranger returns a lost backpack. A diagnosis brings grace. Each thread opens in a first-person voice with a hook strong enough to survive a scroll past, and each thread is a fiction the platform pays for.

AI-generated confession threads clear the 5-million-impression bar cheaply. A creator iterates voice and hook against audience response, posts at peak hours, and lets the recommender handle distribution. The cost is a model subscription. The payout is a share of ad revenue on every view. The 5-million-impression floor favors accounts that already have traffic but does not require original reporting, primary footage, or even an identifiable author.

X has run three enforcement waves against adjacent abuse, and none of them have hit this genre.

In May 2025, X sued 8 named Vietnamese creators and 25 "John Does", alleging they fraudulently drew payouts by posting phony AI-generated content and boosting engagement with bot-driven comments. The complaint named coordinated inauthentic behavior, not the genre itself. In April 2026, X cut payments to "aggregators" who primarily reposted news from other sources and warned that overuse of the BREAKING tag would draw sanctions. On July 16, 2026, X disclosed it had removed millions of posts that purloined already-published content from others. Each action was defensible. Each action was also narrow.

AI-generated first-person melodrama falls outside all three. It is not bot-boosted engagement on a single fraudulent post, the legal theory behind the May 2025 lawsuit. It is not a repost of a news item, the April 2026 target. It is not the verbatim lifting of a copyrighted text, the July 2026 disclosure. It is original, machine-written fiction in a voice the reader cannot distinguish from memoir, paid for by a program whose eligibility criteria make no judgment on the content.

X is now nested under xAI, which is itself part of SpaceX. The platform declined to answer WIRED's questions about the AI slop phenomenon. The same parent that builds the models generating the content also runs the program that pays for it and the legal team that has so far sued adjacent fraud rather than the genre itself.

The next enforcement round will test whether X names the genre. The May 2025 lawsuit showed X will sue AI-content fraud when the fraud is provable and the behavior is coordinated. The April 2026 aggregator cut showed X will redefine the eligible set when the set is being gamed. The July 2026 mass-removal showed X will pull large volumes of content when volume itself is the offense. AI-generated first-person melodrama is none of those. It is a content category the program rewards, at a scale the platform pays for, by a class of creators the platform has not yet named. The marketplace did the rest.