The Plagiarism Machine
Back in the day, as an editor, I was witness in varying degrees to various plagiarism cases. The thing about plagiarism accusations (real ones, not nonsense vexatious lawsuit ones) is, every damn time, over a period of weeks or months, they go like this:
1. Accused is hurt, outraged, appalled. It’s a vendetta. They swear there is no basis at all for this accusation! … Well, there may be some similarities but they had read the plagiarised work as a matter of research, and they swear it must simply have seeped into their consciousness without them realising. … All right, yes, they did lift sections of the plagiarised work, but they were under great financial or personal stress with health problems, it was a one-off situation, and they absolutely pinky-promise hand-on-heart swear they’ve never ever done it before.
2. You find all the times they’ve done it before.
There is no such thing as a one-and-done plagiarist. It is a habitual behaviour, and so is the denial, to the hilt, in the teeth of the evidence, to the point where the accuser feels frankly embarrassed for them.
I have reached the conclusion that in almost all cases, they believe what they’re saying. I don’t think plagiarists sit there scheming to steal other people’s work: I think they persuade themselves (with increasing ease) that the work has become theirs. That a few changes makes it their own, that they haven’t done anything wrong, because they put in work too!
I have (or am certain I have) recently had my second direct professional encounter with someone using AI where they absolutely should not. I am not going to go into details on this because I can’t prove it, but in both cases, I got a shockingly incompetent bit of work with errors that made no sense coming from a human, but were entirely on brand for AI. In both cases, I said that I thought AI had been used. In both cases, the person was asked, and denied it. In both cases, I think they lied.
Plagiarists lie. They lie fluently, habitually, and (very significantly) to themselves. And I see a lot of similarities between your hand-crafted artisan plagiarist and people who use the plagiarism machine. When I see someone insist that ‘crafting’ (amazing verb choice) a GenAI prompt is creative work on a par with writing a novel, such that their AI-generated text is human authored really, it sounds functionally identical to plagiarist self-justification at work. Plagiarists lie, and the plagiarism machine industry is built on lies.
You may have read the story about the $2.4 million deal for a debut novel with film interest that’s fallen apart because of accusations the author used GenAI.
After an editor raised concerns over potential AI use, however, after a process of attempting to validate the manuscript, [the agent] made the decision to pull the novel from sale and cut ties with the author. [The author] told The Bookseller: “Within six hours of the rumours, I lost US, UK, translation, and screen opportunities without being given a proper chance to explain my writing process” … “the story, characters, scenes, humour, structure and voice came from me.”
To be clear, I have no inside knowledge of this sad affair. I do, however, know publishers and agents. I have seen a multinational publisher torch a decades-long relationship with a household-name author over £1500. If you tell me an agent abandoned a $2.4 million deal without making stringent efforts to establish the truth and save their 15%...that is not something that fits my current understanding of the world.
Again, I don’t know the truth here. (Though I am familiar with equivocation, and “the story, characters, scenes, humour, structure and voice came from me” is noticeably not the same thing as “I did not use GenAI”, a statement which isn’t reported in this article.) I will, however, remind you of the saying, “It’s not the crime, it’s the cover-up.” Show me a case of plagiarism and I’ll show you a plagiarist who made it worse for themselves by lying about it. And I see a similar pattern with GenAI use where it should not be.
This is just one case, if high-profile; many will follow. There are going to be a lot of pulled deals, because publishers don’t want uncopyrightable books or social media firestorms. Everyone who follows authors on social media will already have seen numerous cases of flat denial followed by reluctant admission followed by crash-out. And a lot of authors are currently shafting themselves in deeply regrettable and unnecessary ways.
Here’s my advice.
· Don’t lie to your agent. Don’t lie when negotiating legal agreements.
· The best way to avoid accusations of AI use is not to use it. If you do get unjustly accused, as is unquestionably happening, at least you’ll have truth on your side.
· If you insist on using it, don’t lie to yourself. Don’t copy paste ‘research’ or ‘placeholder text’ into your MS while telling yourself you’ll rephrase it later, or pretend that accepting Claude’s ‘edits’ verbatim doesn’t mean the text is now full of slop. You didn’t write it. It’s not yours. A prompt is not work.
· And don’t lie to readers. If you put Human Authored on slop in order to sell it for money, you’re nothing more than a thief. (If you auto-generate fanfic, a thing people create entirely for the fun of it, in order to pretend you wrote it…okay, no, I honestly don’t know what that’s about.)