Surprise. The quiet part got filed in court.
Publishers Weekly walked through newly unsealed material from the Authors Guild case against OpenAI and Microsoft. Executives weren’t guessing about what generative models would do to writers. They were naming it out loud years ago. One 2019 note: they trained GPT-3 on pirated stuff, and “no sharing that.” Jack Clark, then at OpenAI policy, said in 2020 that the better the models got, the more genre-fiction authors would worry about being substituted on Amazon — and that when artists complained, the company would likely ignore them and ship anyway. Another researcher’s stated mission was a machine that would supplant human authors. Complaints about stolen datasets and lost work? “Acceptable economic disruption.” Somewhere in there sits the cheerful line about the death of the reader: machines making slop for more machines.
Been there. I released GOAD, a horror thriller in June. After working on it for 4 years without any kind of AI aid. I feel shattered. I’ve felt the marketplace get noisier but couldn’t prove why.
Then The Atlantic put a face on the next twist. Not just volume. Names. Indie authors finding AI-looking titles glued to their Amazon author pages — same character names, echo covers, fan emails about books they never wrote. One writer’s readers bought the fake, left the kind of review that sticks to a real career, and had to be told afterward they’d been fooled. Amazon often delinks after a complaint. Getting the listing gone for good is harder. Verified authorship at upload would end most of this overnight, says the Alliance of Independent Authors. Instead writers play whack-a-mole with their own bylines while the platform treats a stolen name like a “coincidence.” I have to tell, I haven’t visited my sale page on Amazon since the release. I’m afraid to.
And no shit. It really is happening at scale.
MIT’s Initiative on the Digital Economy covered the volume side: researchers looked at more than fourteen thousand self-published genre books on Amazon from 2023 into 2026. About one in five had substantial AI text by their detector’s cut. The catalog exploded far faster than sales. Revenue per book fell. Generative AI, they write, reshaped the book market without competing on quality at all — displacing revenue and rank through sheer volume.
So the big world is clear. Train on the corpus. Flood the shelf. Wear the name when the algorithm will let you. Call the damage disruption and keep shipping.
Circle is being built as the little-world opposite of that stack.
I don’t know what it gains us, other than a quiet place to exercise craft building with open eyes. Not undetectable. Not anonymous volume. Not a byline that floats free of a person who can answer for the work. Circle is a gated workshop: you request an assignment, you write a real review, you earn credit toward having your own work reviewed, and you do it under a name that means something. When AI is in the mix, disclosure is by role — what the tool did, what the human decided — and a human stays accountable for the judgment. That’s the Springer-shaped idea we’ve already been threading into apply and review: how the tool was used matters more than a magic score that pretends authorship is a percentage.
Honest attention is scarce. Earned credit is how you spend it without lying. If the shelf is crowded with parasite titles and un-owned slop, the answer is not a better detector theater. It’s a place where the person and the role are visible, and the work has to survive other writers looking at it.
We’re not asking the marketplace to become kind overnight. We are building a room where a name still costs something.
Also on the beat
- Detector ≠ authorship: Dartmouth’s paper flagged Provost Santiago Schnell’s op-eds and articles as heavily “AI-written” via Pangram; he says he uses AI for refinement, reviews the final text, and takes responsibility — and that a detector score is not a determination of authorship. https://www.thedartmouth.com/article/2026/09/schnell-ai-writing
- Springer Nature’s risk frame: Updated AI policies for authors, reviewers, and editors stress accountability, transparency, and human oversight — focusing on how AI is used (and when it may not replace judgment) rather than a blunt ban. https://group.springernature.com/us/group/media/press-releases/sn-policy-responsible-ai-use/53209516
- Amazon flood, measured: MIT IDE summary of the Stony Brook–linked study — substantial AI text in ~20% of sampled genre books; catalog growth far outran sales; top AI-heavy titles also leaned hard on distinctive borrowed phrasing. https://ide.mit.edu/insights/ai-assisted-writing-is-it-good-for-authors-and-book-publishing/
- Style without consent: OpenAI is testing ChatGPT Work personalization that pulls writing samples from connected Gmail, Drive, Slack, SharePoint, and Notion so the model matches your tone — useful for ops, messy for who owns the voice. https://basic-tutorials.com/news/chatgpt-work-learns-your-writing-style-openai-tests-access-to-gmail-slack-and-notion/
- Parasite authors, full report: The Atlantic’s deeper cut on AI listings hitchhiking on real indie author pages — reputational harm, complaint-driven takedowns, and why volume genres get hit first. https://www.theatlantic.com/technology/2026/09/ai-authors-impersonating-writers-amazon/688709/
How this was made
AI assisted research and a first draft of this Circle View. A human reviewed, edited, and is accountable for the judgment and the final take. Disclosure by role — not undetectable.
- A score is not a confession → Page 1 Still think a percentage is proof? Read Do AI detectors work on fiction?
- A Detector is NOT a Judge → Page 1 The short version of why a score can’t judge a book: Do AI detectors work on fiction?
- Prove you didn’t use a machine → Page 1 If someone’s asking you to prove a negative, start here: Do AI detectors work on fiction?
- Disclosure that never faces the reader → Page 2 Here’s how we keep it simple in Circle, no formulas: How Circle handles AI
- How did you use AI? → Page 2 We ask once, and your word does the rest: How Circle handles AI
- Put it on the Imprint Page → Page 2 Why I’d rather skip the label tiers altogether: How Circle handles AI
- Watermarks don’t sign your name → Page 1 Watermarks, detectors, same problem: Do AI detectors work on fiction?
- They knew the shelf would get crowded → Page 2 Slop is the problem, not the tool. Here’s where Circle stands: How Circle handles AI
- Label the voice → Page 2 How we handle AI in Circle, without the labels: How Circle handles AI
- Byline is no Costume → Page 2 Your name, your word, your work. That’s the deal in Circle: How Circle handles AI
