The Drey Dossier

The Drey Dossier

Playback speed
×
Share post
Share post at current time
0:00
/
Transcript

The Shy Girl AI Scandal Is Way Worse Than You Think

The Drey Dossier's avatar
The Drey Dossier
Apr 02, 2026

The weeds are already being decided. That’s what I keep coming back to after everything I’ve found since Part I. I wrote about flowers getting diagnosed as weeds, about systems that were never built to protect the people they judge, and I thought I was describing where we were headed. I was describing where we already are. Because in the weeks since that piece went out, every thread I’ve pulled has come apart in my hands, and the picture that keeps emerging is so much worse than I thought.

We are woefully unprepared for any of this. And everything I found just keeps proving it.

If you read Part I, you know the basics. Mia Ballard self-published a horror novel called Shy Girl, it found its audience, and Hachette, one of the five largest publishers in the world, acquired it. Then the internet accused her of writing it with AI, a detection company called Pangram scanned the book and returned a number of 78 percent AI-generated, the New York Times ran the story, and Hachette dropped her the same day. This is the first time a Big Five publisher has ever pulled a commercially published book over AI allegations, and it is setting the precedent for how every accusation like this gets handled from here.

I genuinely do not care whether Mia Ballard used AI. That question, as consuming as it has been for the internet, is beside every point I am about to make. I need you to let go of it so you can actually hear what I’m saying.

Every institution involved in ending this woman’s career had the power to do better and chose not to. Nobody did their job. And the only person who paid for it was Mia. Since Part I, I have new confessions, a sales pipeline disguised as journalism, and I’ve been in touch with Mia Ballard herself. So let me show you how this keeps unraveling.

The Drey Dossier is reader-supported. To support my work, consider becoming a paid subscriber (:

Where It Starts

After my first piece went up, I came across a blog post from Thad McIlroy, the publishing consultant the New York Times cited as one of their two sources of evidence against Mia, and reading it was like watching someone carefully lay out all the evidence against themselves and somehow not notice. Five days after the Times piece ran, he published this post on his website detailing his entire involvement in the story, and the very first thing he wants you to know is that he brought the scoop to the Times, worked with them to develop it, and is already frustrated that the published article obscured how central he was to all of it.

That’s a revealing thing to lead with. But it gets more revealing when he tells you how he first heard about the book.

He writes that he learned about Shy Girl at the beginning of February from Asia Laird, whom he describes as recently hired at Pangram. When I read that name I wanted to know who she was, so I looked her up, and what I found reframed the entire story. Her LinkedIn says Sales and Business Development at Pangram. ZoomInfo lists her as Founding Account Executive. She is one of Pangram’s first hires, and her job is to sell their product, and their product is the AI detection tool whose scan produced the 78 percent number that the entire world has been repeating ever since.

So the person who surfaced this story to the man who brought it to the New York Times was a sales employee at the company whose product became the key evidence. That is where the whole thing starts.

McIlroy also writes that he’s known Pangram’s CEO, Max Spero, “for a couple of years” and has “a lot of faith in the seriousness of his technical approach,” which matters because his consulting firm works with publishers and with the tech companies that sell products to publishers, meaning this was a pre-existing professional relationship between a detection company and a consultant whose entire business depends on exactly these kinds of tools.

And here is what makes the timing so interesting. On January 23rd, weeks before Asia Laird brought the story to McIlroy, Max Spero had already posted Pangram’s scan results publicly on X: 78 percent AI detected. And right there in the preview text of the tweet, visible without clicking anything, you can see OceanofPDF.com, a well-known book piracy site. The URL is just sitting there, in the preview of the tweet that the New York Times later linked to in their own article, and somehow nobody looked at it closely enough to notice. I did, in Part I, and as far as I can tell I was the first person to report it.

That tweet got 2,125 views and 17 likes, and nothing happened. Then, a week to ten days later, the company’s sales employee brought the story to McIlroy. I’m not saying anyone sat in a room and planned this. I’m saying the tweet didn’t work, and then the story ended up in the New York Times. You can call that a coincidence. I would call it a pipeline.

The Copy Nobody Paid For

This is where the sourcing starts to collapse in a way that, once you see it, you cannot unsee.

McIlroy describes trying to buy the UK edition online and getting declined because of Hachette’s territorial sales restrictions, which is standard publishing practice, fair enough. But then he writes that he asked a friend in Austria to find him a DRM-free copy and email it to him. For anyone unfamiliar, DRM is the copy protection publishers put on ebooks to prevent unauthorized distribution, and every legitimate purchase comes with it. Asking for a DRM-free copy means asking for a version stripped of those protections, and for a commercially published ebook, that version almost always only exists through piracy.

He says he got a file, converted it to a .docx, stripped out bonus material from the UK edition, and sent it to Pangram, who came back overnight with 78.4 percent AI-generated, essentially the same number Spero’s pirated scan produced. Two supposedly different copies run at different times, same result, and we only have proof one of them exists: the pirated one.

I emailed McIlroy directly, identified myself, linked my previous reporting, and asked him about the copy, and his response was extensive in a way that somehow made the picture less clear, not more. His friend had first sent a file he couldn’t open because of what he called “some strange encryption,” which is almost certainly just DRM, meaning the first file was real, with the protections a legitimate purchase is supposed to have, and McIlroy couldn’t use it. A second file from a different source worked, but he told me he wasn’t sure which retailer it came from. His friend told him it was purchased, and McIlroy took him at his word.

In his blog post he asked specifically for a DRM-free copy, and in his email to me he said the whole point was to avoid piracy, but the request itself described the exact thing that usually only exists through piracy. Every direction this man sends me in raises more questions than the last.

And then he told me something I had to read three times. He does not know for certain whether the copy his colleague sent was Mia Ballard’s self-published original or the Hachette UK version. He has never checked.

But in his blog post, he describes stripping out material that only exists in the UK edition. If you are removing content specific to a particular version, you know which version you have. He later obtained both versions and compared them side by side, and he still cannot tell me which one he built the investigation around.

Every person involved in building a case about the integrity of creative work could not be bothered to simply buy the book.

Five Weeks and One Day

McIlroy reached out to John Maher, a former editor of his at Publishers Weekly who had moved to the New York Times Book Review, and Maher assigned reporter Alexandra Alter. What followed, by McIlroy’s own account, was five weeks of collaborative work: emails, phone calls, answering every question she had. He writes that Maher assured him the reporting would make it clear that McIlroy had brought the story to the Times, so he was already negotiating his credit before the investigation was finished. And he kept pushing them to publish faster, worried that things would drag past the US release date or that the Times would lose interest. The urgency was always about the window.

He also says he would have liked to interview Mia Ballard, but that going through Hachette would tip them off and compromise the scoop, so he made a choice, and the choice was the story over giving her a chance to respond.

On March 19th, the Times contacted Hachette, and Hachette pulled the book the same day. Mia Ballard confirmed to me directly that the New York Times first contacted her on March 19th, the day the story ran.

Thad McIlroy had five weeks. Mia Ballard got the day of.

And then his blog post does something I genuinely was not expecting. After telling you he chose not to contact Mia, after telling you he pushed the Times to publish faster, he writes a section lamenting that Hachette threw her under the bus, asking why the reporter didn’t insist on access to Ballard before running the story. The dissonance is staggering. He is wondering aloud why someone else didn’t try harder to reach her when he is the person who made the deliberate choice to leave her out, who said in his own words that reaching her would compromise the story, and who is now writing about it as though someone else dropped the ball.

And the contradictions just keep compounding. He writes that detection reports should be used “only for guidance, not as proof of guilt” and that “there has to be an underlying presumption of innocence,” principles that apparently did not apply when it was his evidence being used to end Mia Ballard’s career without any presumption of innocence at all. He used an AI-generated summary of a three-hour YouTube video about AI detection to build his case about AI detection, which I know because I watched the entire thing before I made mine, and it took three hours. He investigated whether Ballard’s self-publishing imprint was connected to the Church of Scientology, confirmed it wasn’t, and left it in the blog anyway.

And he closes calculating the return on his involvement, writing that while it’s great to be quoted in the New York Times, he is “distraught” that he received “not much more than a passing mention” and that “there’s near-zero value to me from my involvement.”

Near-zero value. To him. From his involvement. Mia Ballard lost her book deal, her UK publication, her US publication, her social media presence, and according to her own words, her mental health. And this man is calculating what the experience was worth for him personally.

One Rule

I kept digging into McIlroy’s background after reading all of this, and what I found made the whole story land differently.

In July 2024, about eight months before any of this, McIlroy published The AI Revolution in Book Publishing, and Ingram, the biggest book distributor in the country, flagged it and pulled it from distribution after their automated AI detection filters identified it as AI-generated. McIlroy says about one percent of the book was touched by AI, in sections where he demonstrates the tools for the reader, though he also used AI to translate the book into 31 languages and generate six AI-narrated audiobook editions in three different English accents, so whether “one percent” captures the full picture is a question I’ll leave with you.

What matters is what happened next. He posted about it on LinkedIn, and within hours an Ingram executive named Margaret Harrison commented personally offering to look into it, and by Monday the book was back on shelves. The New Publishing Standard called it out for exactly what it was: one rule for important people and another for everyone else, given that the normal appeals process at Ingram takes up to 14 business days and, according to authors on Reddit, can stretch to months.

And after it was resolved, he said on the record in Publishers Weekly that he wondered how many authors might find themselves in similar situations and questioned whether the system was robust enough to protect them.

Eight months later, he put a debut author in exactly that situation.

Do we think Mia Ballard is going to get a LinkedIn comment from a Hachette executive that fixes everything by Monday?

The Diagnosis

I said in Part I that Hachette will litigate copyright when Google is on the other side of the table but won’t defend their own author’s intellectual property when it’s used against her. I said the White House just sent Congress a framework designed to make sure no one like Mia ever gets protected. I said the tools carrying documented racial and linguistic bias are being treated as evidence in career-ending decisions. All of that still stands, and none of it has changed.

What has changed is that I now know how the diagnosis gets made. A sales employee brings a tip to a consultant with professional ties to her company, and the consultant sources a copy of the book he cannot verify, through a process that by his own description mirrors piracy, and sends it to that same company for analysis. The company produces a number, the consultant spends five weeks shaping the story with the New York Times without once contacting the author, and then the story runs and the book is pulled the same day. And the consultant closes his blog post calculating whether the experience was worth it for him.

That is the pipeline. That is how flowers get diagnosed as weeds.

And if it can happen to her, with this little rigor, built on evidence this shaky, laundered through institutions this careless, it can happen to anyone. Which is why I covered it, and why I’m still here.

Mia and I have spoken multiple times. She knows what this reporting is about. She knows I am here to stand up for the standard that should exist so that what happened to her cannot happen to the next person without anybody answering for it.

I don’t have a newsroom. I don’t have a consulting firm. I don’t have a title at a trade publication.

What I have is a promise. That as long as the people at the top keep building systems that protect themselves at the expense of everyone else, I will be here, pulling the threads they don’t want pulled, asking the questions they hope nobody asks, and publishing what I find whether they like it or not.

Today it’s Mia Ballard. Tomorrow it could be you. And I’ll be here for that too.

SOURCES CITED

Discussion about this video

User's avatar

Ready for more?

© 2026 Audrey Henson · Privacy ∙ Terms ∙ Collection notice
Start your SubstackGet the app
Substack is the home for great culture