63% of Amazon Religious Books Likely AI-Written, Study Shows
In brief
- Originality.ai flagged 1,272 of 2,034 Amazon religious books as likely AI-written across 14 categories
- Witchcraft books hit 78% AI rate, followed by Hinduism at 76% and Taoism at 74%
- Amazon relies on author self-identification rather than enforcing AI-detection tools
- AI detection tools show conflicting results, raising accuracy and reliability concerns
The Breakdown Across Belief Systems
Witchcraft had the highest rate, with 78% of books sampled from Amazon's Wicca, Witchcraft & Paganism category classified as likely AI-written. Hinduism ranked second at 76%, followed by Taoism at 74%. On the lower end, Mormonism came in at 42%, atheism at 40%, and Satanism at 22%.
The witchcraft category presented a particular pattern. Books covered subjects including healing crystals, herbal remedies, and cleansing negative energies. These titles appear to target readers seeking wellness solutions outside conventional medicine, a market segment where AI-generated content may proliferate with minimal scrutiny.
How the Detection Works—And Its Limits
Researchers analyzed book descriptions, author bios, and samples, classifying text that scored 50 or higher as likely AI. But AI detection itself remains unreliable. In October 2024, a test of four AI detectors produced conflicting results when analyzing the U.S. Declaration of Independence. ZeroGPT found the text to be 97.93% AI-generated, while QuillBot found it entirely human-written and GPTZero gave it an 89% probability of being human-written.
Real-world consequences have followed. In March, Colombia's Supreme Court rejected a legal filing after detectors flagged it as AI-generated. Yet when an attorney submitted the court's ruling to the same detector, it was rated 93% AI-generated. The irony is stark: a human-written legal document scored as AI, while actual AI-generated work scored as human.
Amazon's Hands-Off Approach
Amazon uses the honour system for authors to self-identify their work as AI-written. No mandatory detection. No enforcement. The platform appears aware of the scale of the problem. In August, an investigation tracked a shipment of rare books to an Amazon facility in Las Vegas, where workers cut off their bindings and scanned the pages for AI training data. Amazon confirmed that it purchases books through commercial channels for the operation.
This creates an odd dynamic: Amazon profits from AI training on human-authored books while allowing AI-generated books to flood its store under minimal disclosure. The religious book category—where belief, authority, and lived experience matter—becomes a testing ground for how little friction a platform can tolerate before stepping in.
