Security researcher Michal Zalewski has documented a troubling pattern in AI-generated content: roughly one hundred fifty nearly identical machine-written books flooding Amazon's non-fiction categories, many ranking as bestsellers. While individually indistinguishable from legitimate work, the covers expose the real problem—recurring design elements across dozens of titles, from roaring dinosaurs positioned identically in the top left corner to red-and-white rockets and golden retrievers appearing in predictable clusters. This pattern reveals how large language models resort to the same set of mannerisms regardless of the prompt. Unlike human authors, who develop distinct voices, AI systems gravitate toward statistically probable responses. Zalewski argues this makes AI writing detectable not through individual quirks, but through repetitive patterns at scale. As artificial intelligence makes content production far cheaper than human evaluation, he warns, this threatens the viability of meaningful online discourse.

Listen to this story

Hear this and more stories in a personalized audio briefing.

Open The Chonkerton