In other words, I wasn't suggesting Boucher would have legal issues. I was suggesting he might have PR issues if his work was perceived as Ai-generated. (He was being called a fraudster, not a copyright infringer.)

Wasn't he acknowledging upfront that the works were AI-generated?

Quote from: Post-Crisis D on September 03, 2024, 09:36:34 AM

Quote from: Bill Hiatt on September 03, 2024, 06:26:42 AM

Semi-solid ground, assuming his own statements are accurate. But if challenged, could he prove that the ideas came from him? The short answer is yes, but he'd need to produce the prompts and responses for comparison to the finished work. If he's saved all the supporting materials, he's in good shape. If not, his position could be problematic down the road.

Ideas are not protected by copyright so it wouldn't really matter where the ideas came from.

I was responding tot he part of his claim that the ideas are his, and he's only using AI to help implement those ideas. That's fine--as long as the ideas really are his. The issue here is whether there's a copyright infringement as much as it is how much the human author is actually contributing to the finished product.

Again, it doesn't matter whether the ideas are his or came from the AI tool or came from somewhere else.

The expression of the idea is what is protected by copyright.  And, if the AI reproduces copyrighted material and he ends up using it in his works, then there could be copyright infringement.

If I understand the current U.S. Copyright Office's interpretation of things, AI generated content cannot be protected by copyright.  I think it is considered public domain, excepting cases where it may reproduce copyrighted material.  So, the only portion that would be protected by copyright are those portions written by a human author.

Of course, anyone other than the author would have no idea what part of the work is technically public domain and what part is protected by copyright.

And that begs the question whether authors (and other creatives) should have to disclose (and perhaps publish?) exactly what is and is not AI-generated.  Granted, the copyright application (U.S.) asks you to limit/describe the copyrighted portion of your work, but how many authors follow that precisely?

For example, Frankenstein is public domain.  Anyone can do whatever they want with it.  Maybe Frankenstein is a winning race horse that dies but nobody knows and Dr. Frankenstein works to bring it back to life so it can keep winning races because the horse's owner, which may be Dr. Frankenstein himself, is in debt and just needs to win two more races to pay everything off.  Now, if you write that, I can't use anything except what was public domain from the original.  Fortunately, I can get a copy of the original Frankenstein novel and create my derivative work based on that.

In the case of AI works, where can I look up the original public domain work if an author uses AI to create his/her work?  Should people be able to see that somewhere?

And it's a darn good thing that ideas are not copyrightable, because I just discovered that a plot thing I used in a recent book was also used in a book I'd read forty years ago (and had totally forgotten). Obviously, the circumstances and details were different, but basically the idea was the same.

There aren't all that many ideas in genre, and that's the way genre readers like it. More of the same, please--but not exactly the same.

Semi-solid ground, assuming his own statements are accurate. But if challenged, could he prove that the ideas came from him? The short answer is yes, but he'd need to produce the prompts and responses for comparison to the finished work. If he's saved all the supporting materials, he's in good shape. If not, his position could be problematic down the road.

What is interesting is he says he doesn't sell his books on Amazon.

Just avoiding a gatekeeper he knows will close the gate