Amazon's review rules ban conduct, not how text was made

The writer's problem

An author or small publisher watching reviews accumulate on their book's Amazon page needs to know which reviews the retailer will actually act on: an unfavorable review is not by itself a policy violation, but a compensated or fabricated one is. As retrieved 16 September 2026, Amazon's own Community Guidelines set out what counts as prohibited review conduct on its platform.

What the documents show

The guidelines state reviewers may not create, edit, or remove a review in exchange for compensation, and that this applies even if a seller offers a refund to remove or change a negative review. They separately prohibit posting a review on a seller's own product, on a competitor's product, or on the product of anyone with a personal connection to the reviewer, and they require anyone with a financial or personal connection to a brand, seller, author or artist to disclose that connection when posting content about it. The guidelines describe reporting channels, including a dedicated form for reviewers offered compensation to write, edit or remove a review. A separate Amazon-published page on fake review brokers states the company's own enforcement figures: it says Amazon blocked more than 250 million suspected fake reviews in 2023 and took legal action against more than 150 bad actors that year, and that brokers increasingly use automation and artificial intelligence to solicit fraudulent reviews at scale. Neither document names AI-generated review text as a distinct category; both are written around conduct, compensation, and disclosure, so a review's underlying rule violation is the same whether the text was typed by the reviewer or produced with an AI writing tool.

The editorial choice

An editor explaining review manipulation to an author should present the rules as conduct-based rather than technology-based: what makes a review prohibited is the undisclosed exchange or the fabricated origin, not the software used to draft it. This is an editorial framing the sources support rather than state outright, since Amazon's guidelines and its enforcement page discuss automation only in the context of brokers scaling abuse, not in the context of an individual reviewer using a writing tool to phrase a genuine opinion.

What stays with the author

The guidelines do not tell an author how to prove a suspicious review was compensated, and Amazon's own figures on blocked reviews and legal actions are the company's self-reported enforcement record, not an independently audited total available for a reader to verify. Reporting suspected compensation through Amazon's own channel, rather than assuming the platform will detect every violation automatically, remains the step available to an author who suspects manipulation.

• Is the objection to a review its content, or a specific compensation or disclosure violation named in the guidelines?
• Are enforcement figures the platform's own self-reported numbers, or independently verified?
• Does a policy address conduct in a way that applies regardless of what tool produced the review's text?

Amazon's own guidelines and its enforcement statements describe a conduct-based prohibition on compensated, fabricated and undisclosed-conflict reviews; they do not, on their own terms, single out AI-generated text as a separate category of violation.