Authors Guild tells publishers: detectors alone can't justify cancellation
The new guidelines follow a similar set the Guild issued to authors in May, aimed at building "mutual trust" on both sides.

This is the Guild finally putting in writing what every author with a shaky "AI score" has felt in their gut: one algorithm's guess should never be the whole case against you. I like that it doesn't let authors off the hook either — transparency cuts both ways here. But guidelines are only as good as what you do with them, so don't just nod along and move on.
Here's the fear every author's been quietly carrying: an algorithm flags your manuscript, nobody questions the algorithm, and your contract is gone. The Authors Guild just put out guidelines for publishers that take direct aim at that scenario.
The core of it is simple. "No publisher, agent, or other person should rely solely on an AI detector to determine if a book contains AI text," the Guild states, adding that even the best detectors "are not always accurate." The Guild goes further, saying it "objects to the use of AI detectors alone by publishers, editors, and agents to accuse a writer of using AI to generate text, much less to cancel a contract." Their reasoning is blunt: "One false accusation can ruin an author's career." They're right.
This follows a similar set of guidelines the Guild issued to authors back in May — the goal of both documents, the Guild says, is "to ensure mutual trust between authors and publishers." That matters. This isn't a one-sided author wish list; it asks something of publishers too.
Specifically, the Guild wants publishers to be "fully transparent about when and why they use AI detectors" and to put that in a written AI policy. If a publisher is sending your manuscript off to a third-party detection service, the Guild says the publisher "should know how the service stores and uses submitted text, should confirm that the text will not be retained or used for training, and should obtain the author's permission." That last part is the one to remember: your permission. Not a formality buried in a submissions portal — actual, informed consent before your pages go anywhere.
The Guild also drew a hard line on contract cancellations generally: "no publisher should ever cancel a contract due to AI use unless the author violated a signed contract or policy that clearly prohibits the particular use and it is a material term of the agreement." Vague suspicion isn't grounds. A clear, signed, specific policy violation is.
And on the creepier end of the AI-policing trend — tools that track or log your writing process to "prove" you wrote your own book — the Guild opposes making that mandatory. Though if you want to use that kind of tracking voluntarily, as a shield in case you're ever accused, the Guild says that's your call to make.
What to do this week: if you don't already know whether your publisher has a written AI policy, ask. Ask specifically whether they use AI detectors, which service, and whether they require your permission before submitting your manuscript to one. Get the answer in writing. If a publisher can't tell you their policy exists, that's your answer too.
Why it matters to you: If your publisher ever flags your manuscript with an AI detector, the Guild's position is now on record that detector alone can't be the reason your contract gets cancelled.
This story ran in the Coral Wire edition of Sat, Oct 17 2026, when the Coral Index stood at 53/100 (Mixed signals).
Coverage — 1 source
Every outlet found reporting this story.
- Authors Guild Issues AI Guidelines for Publisherspublishersweekly.com
Also in this edition
- Faber pulls Sally Rooney titles from UK sale over terrorism lawLegal — Normal People sold a million copies in the UK, and it's still coming off the shelf
- Agent explains the foreign rights trade-off before you signLegal — Keep the rights and the money comes direct, but only if your agency actually has a foreign rights desk
- Orélien admits old plagiarism, denies AI use on new novelLegal — La Presse's investigation into his 2012-2014 work got him dropped from Goncourt Prize contention, but the AI accusation stands separate and denied.
- Bramble lands two seven-and-eight-figure fantasy deals this monthMarket — Devney Perry's City of Saint series went for eight figures and Sarah Beth Durst's Spellshop extension for seven, both at Bramble.
- Authors and industry groups launch global ethical AI charterLegal — The Society of Authors says its own position on generative AI "remains unchanged" despite the backlash the charter has drawn.
- Substack's Pangram AI-detector sparks writer civil warAI — A seven-figure book deal, a Hachette pull, and a lost NYT column show publishing can't agree what the AI problem even is
- YouTube in talks to launch audiobooks next yearPlatform — The Wall Street Journal reports YouTube may bundle the feature into its $16-a-month Premium tier.