.jpg "Amazon’s AI infrastructure spending to hit 0bn this year")
For 6 months, my husband convinced everyone that I had stolen millions from the company I built, including our own daughter. As I sat in federal court, certain I was about to lose my freedom, the doors suddenly opened. My 9-year-old son walked in carrying an old blue backpack, looked straight at the judge, and said, “Your Honor. I know who framed my mom. They’re sitting in this room.”
So begins a thread on X that appeared in my feed one day. I knew that the first-person narrator probably didn’t exist, and following this thread would draw me into a mawkish tale where good triumphs over evil. Indeed, the author is not a mother on trial for theft, but a self-identified 21-year-old stock trader in Nigeria, who uses AI to generate stories—and gets paid for it by X.
I found this out when I began wondering about the proliferation of these compact melodramas of redemption in my feed. It’s probably my own fault; I sometimes idly click on these cheesy stories—you know, just to see what happens next—and the algorithm took notice. In this case, the author is one of countless beneficiaries of X’s Creative Revenue Sharing program, which pays for posts that boost engagement on the platform. To join, a creator must become a Premium X user, have at least 500 verified followers, and have drawn a minimum of 5 million impressions in the last three-month period. The stated goal is to encourage “authentic, high-quality content,” and some of the money surely goes to creators doing just that.
But paying for posts that go viral makes the program ripe for exploitation. Some creators in the program engage in “rage baiting,” using outrageous content to draw comments and views. In May 2025, X sued 8 Vietnamese creators and 25 “John Does,” alleging that they fraudulently drew money from the program by posting phony AI-generated content and then boosting engagement by bot-driven comments. In April, X moved to cut the payments of “aggregators” who primarily posted news from other sources, and threatened sanctions against those who overused the BREAKING tag to draw eyeballs. On July 16 of this year, X revealed that it had removed millions of posts that purloined already-published content posted by others.
It’s not clear how X regards the AI-generated story slop that has infected my feed. (The company, now part of xAI—which is itself now part of SpaceX—would not answer my questions about this.) Most of the stories are falsely presented as true first-person narratives, though it’s pretty obvious that they are fabricated. In many cases, the creators describe themselves in their profile as storytellers. It’s hard to prove that the actual authors are AI models, but the vast majority bear the signs. The posts are manipulative and dumb and anything but authentic.
Rob Cover, a professor at Australia’s RMIT Digital Ethnography Research Centre who studies AI-generated content, broke down the structure of these posts for me in an email. It’s as formalized as a sonnet. First comes a rage-baiting setup, “in which an innocent person is wrongfully fired, evicted, scammed, loses their rightful place on public transport, or is the subject of judicial overreach.” This is followed by a turnaround: “The underdog is redeemed while the adversary is defeated, humiliated, and shamed (often as a representation of a public shaming),” says Cover.
In many of these tales, narrators are in possession of sealed documents that turn the table on tormentors, often in divorce disputes or the execution of a relative’s estate. Or it could be something as simple as an airline passenger who fiercely retains their assigned seat when rude moochers lay claim to it—and the adversary suffers a walk of shame to the back of the plane.
