Actually, I had a rather significant Net incident with some large media entity based in Canada, but it was a number of years ago and right now I only remember that I didn’t know just how large that company was until it got serious. In fact, I can’t even remember how that matter was resolved. I think I was under contract with Time-Life then and they took over the matter.
Now that is scary!
But it is scary because that takes us into very deep forest that involves who might be watching.
I’ve had to do some research on this issue and some of the results I grabbed were quite a surprise.
BUT, that was before that Gemini thingy started inviting itself onto my screen and I think them Microsoft folks have their own style of some “Let me think for you.” type program, too.
In fact, I think even DuckDuckGo has some sort of such on their search page results. Or one can click for it to show; I forget as I type this. I’ve got too many browsers now and my brain can’t keep up anymore.
Me? No. I pay a lot of attention to copyright issues generally, and am somewhat familiar with how the current incarnation of LLMs works, but I don’t have any information about formal certifications.
I’d like to see both the authors and publishers involved in cases like this stand up and fight, rather than curling into a defensive crouch. A false accusation of AI use is clearly defamatory, and in a defamation action the burden of proving the accusation true falls on the person making it. (Not a lawyer, not legal advice in any jurisdiction.) It’s disappointing that they haven’t, but of course it’s impossible to know whether the reason is that the accusation is true, or that litigation is expensive and risky.
If I were some twenty years younger I could kick out one heck of a story on this.
I just used AI to check on AI and the Monster AI called Gemini informs me that AI laws and regulations are all mixed up and confused and I would, as a human, describe what exists as laws and regulations around the planet as a big-big-big plate of spaghetti. Maybe a serving tray of spaghetti.
Oh yes, and Gemini pointed me to that most reliable of law firms, UTube! Now ain’t that fun!!?
In fact, my weird brain just remembered something.
I’m sure I remember seeing a website’s rule that no AI generated stuff was allowed to be posted.
I don’t remember when or where I remember that being, but I am sure I have the key point correct.
BUT, if even just checking GIL for information, or BIL, or any other search engine . . . well, it seems that rule about “No AI Stuff” could be really stretched and anything could be given a label of AI.
Even stating you saw a headline and . . .
Sure seems you’d have to write one very long explanation of what you define as AI generated.
You younger folks out there in the Net World sure do have some trouble up ahead!
Well, unless you are going for a law degree. Then you’ll get rich!!
And you are on the cusp of the real problem… “Not Invented Here Syndrome” and “Not My Idea Disease” are not new. “My way or go away” isn’t either. Have you read about how the academics of the 19th century got after a few drinks and the knives came out? The tools are different, the details are different, but the underlying tribalism is not.
AI is just the latest whipping-boy. Oh, I fully grant that OpenAI and the other major AI companies do themselves no favors by the way they beg/borrow/cheat/steal their training data… But it’s just the next generation of Plagiarism!
You read my mind, as I, too, would like to see accused authors stand up and fight. (I bet there are lawyers itching to take on the right case who would happily do so on a contingency fee basis. Were I still practicing, I’d be one.)
In my view, falsely accusing someone of using generative AI to write a novel is defamatory (libel if written, and online posts count; slander if spoken). But a defamatory statement alone isn’t actionable; there are additional elements to prove.
As to the burden, it depends on which country’s law applies.
In the UK and most of Canada, you’re right: falsity is presumed, and the accuser must prove the accusation true.
But in the U.S., where I practiced and taught media law in the pre-AI era, it works differently. Because an accusation like this almost certainly involves a matter of public concern, the burden generally falls on the author, as plaintiff, to prove a false statement of fact. That alone could sink a case built on “this reads like AI,” which may well be protected opinion.
Unlike in the UK or most of Canada, the U.S. First Amendment requires the author to prove fault. For an author who is a private figure, that means proving at least negligence; for a public figure (which a well-known novelist could be), it means actual malice: that the accuser knew the accusation was false or seriously doubted it (aka reckless disregard of the truth), shown by clear and convincing evidence.
So in the U.S., we’re off to the races on a different question: not how the accuser proves the author used generative AI, as it would be in the UK or Canada, but how the author proves a negative. Scrivener snapshots, anyone? (And yes, I know that even careful snapshotting of one’s writing progress might not be enough.)
As for why so few fight back, in the U.S. about 40 states have anti-SLAPP statutes. They let a defendant who is sued over speech on a public issue seek early dismissal and recover attorney’s fees from the plaintiff. Egad.
Assuming the author gets past all that, the damages would be interesting, too. At least for a signed author, immediate financial losses could probably be established, and perhaps future loss of earnings.
One more thought. While a publishing agreement likely gives the publisher a way out, such as a morals clause or, increasingly, a warranty against AI use, I’m not so sure such contracts typically shield the publisher from liability for publicly accusing an author of using generative AI. Unless the author consented to that, I don’t see why an author couldn’t sue the publisher for defamation, subject to the same hurdles.
(Not legal advice.)
This is an issue I started contemplating about 20 years ago because of certain duties on the Internet for an entity that was new back then, and still going strong now, but with a very different set of platforms.
As we had to deal with Japanese, ROK, PRC, ROC nationals as the primary concerns with regard social media arguments; and then a smattering of others from a number of other nations - - - well, obviously one’s mind starts to wander into why can’t there just be one international law-type arrangement. But it was such a stupid idea, it would only once in a bit be brought up in the back “rooms” where the admin and mods mixed and had time to yak about non-work stuff.
Odd thing is that even before that, by a number of years, when I was on active duty there was this thing called SOFA and it basically allowed for little islands of area inside a sovereign nation where that sovereign nation’s laws didn’t apply; sort of. And that “sort of” was no small matter in itself. In fact, that SOFA umbrella could actually be carried out one of the gates and apply in a village or town, or even a large city. Very weird stuff.
But back to these days, one biggie now seems to be where a site’s server is located. And that can get mighty weird.
I’m a member of the ISOC and not active for many years after a horrid battle with the then president of the organization over selling a big chunk of something, which some of you might remember (they eventually were able to sell that, though [so in the end I lost]) BUT we used to have some mighty heated discussions about how to organize some sort of international treaty, or some such type of regulatory entity; but that is one difficult-difficult-difficult task and . . . well, I can’t go too deeply into that here (I don’t think) - - - I think the way we (Net Citizens) are heading now with this whole Internet is bad for all human culture, not just any one culture/nation. Everyone. Truth is the “bad” is already showing the signs and it really is scary.
Guess I should be careful here, yes? I may be heading into territory the staff/admin may not approve.
Dain, thanks - I think you’re right on the hard part. If a publisher has already decided to drop someone, a timestamp won’t talk them out of it, and fighting that can eat years you’d rather spend writing.
Where I still see a use is the quieter cases: not a nine-year inquisition after the fact, but keeping a trail as you go so you aren’t empty-handed if someone later waves an AI-detector score or a metadata claim at you. Vincent’s later note that Orélien couldn’t show any draft predating the final one is exactly the hole an outside date on earlier files is meant to shrink. It still doesn’t prove who wrote the words, and it isn’t a shield against a bad-faith partner. It’s just one dated fact you don’t control, sitting next to snapshots and a writing group.
You’re also right that it only helps if you start before you need it. That’s the bit I care about for Scrivener folks more than any courtroom fantasy.
It’s also really terrible practice. Failed hard drives and computer viruses existed long before AI. Unfortunate editorial choices that need to be reverted predate not only AI, but computers. Having a routine backup strategy for critical data is a basic survival skill for a writer, AND helps fight accusations of AI use.
Not only that but what really confuses me about this is the fact that having multiple drafts with previous versions (physical, digital, or both) is kind of just a natural result of the writing process. Anyone with work that is of publishable quality should by default have these, unless they’re purposefully discarding them beyond recovery. It’s literally just part of the writing process.
I think this is how publishers/agencies get suspicious. If you have a final product but nothing leading up to it, then how and why?
From what I have read, the dead giveaway is inability to talk about “your” work.
Publishers and agents talk to a lot of writers. They know how real writers respond to questions like “why did you make this choice?” or “what if we changed [aspect of setting]?” It isn’t even the exact answer, which will vary between authors and between books, it’s the process leading to the answer.
Even the most formulaic, least introspective writer on earth still spends a lot of time immersed in and thinking about their own words.
I was asked recently about a line I used in a submission. The conversation went something like this:
– It’s a near quote from David Ormsby-Gore.
– Who?
– The bloke who was hot for Jackie O after JFK was shot.
– That was Onassis.
– Before him.
– A Brit?
– Americans don’t use hyphens. The whole dashes thing confuses them.
– So why this line?
– Because it fits Protag One’s character, and links to Protag Two’s suppression of reality… and the realisation hits hard. It’s also funny, in a dark way.
– It was sad.
– That’s the point.
I really wanted to add: “I couldn’t believe it when ChatGPT suggested it,” but I fear my sarcasm would have been taken too literally.
But your “Why did you make this choice?” point should be the crux of this proving nonsense. I spent about two years toing and froing between the use of “bright” and “pale” in an opening sentence. I could do a TED talk on it. If a writer can’t bore the pants off someone about these minutiae, then I’d wager that’s a definite signal of something.