If a user generates a short story using AI, it has no copyright that the user can claim. Anyone can copy the story. If the user claimed they owned the copyright, one day it could catch up with them?
However, if the user rewrote the story in their own hand, does they then own the copyright? Surely yes as no-one knew the AI story existed ever anyway.
If it is yes, how much has to be rewritten for the ownership of the copyright to move from AI (no copyright) to human?
At this time (Oct 2024) this is still an open legal question with AI.
The problem is that AI learns from example writing, and it has been found that sometimes, AI will inadvertently quote some of its learning sources verbatim, and that quote is copyrighted.
Which would mean you cannot use it. Even the AI cannot use it! And AI companies are currently being sued by the copyright holders for presenting their work as their own "original work".
I agree with them, and the idea that this was inadvertent and they (the AI coders) did not expect it to happen is not an excuse; their AI is making money from copyrighted works.
They did not expect it to happen because they train on literally millions of works; and statistically it should not happen, but a hallmark of AI is we can know exactly how the individual parts work, without understanding how the millions of parts working together actually work.
And it turns out, sometimes because it was trained on this copyrighted passage, that passage turns out to be the most likely thing to spit out, verbatim, in response to a prompt.
Anyway, the law moves slow, and it is going to be a long time before this gets sorted out.
As to whether rewriting the story would work, probably yes, as long as you don't use any passages generated by the AI verbatim (because you think they sound so good).
But plots cannot be copyrighted. Character names can be trademarked (I think all the Harry Potter character names and places (like Hogwarts) are trademarked).
So sure, you can use AI to spark your imagination.
Personally, I have found AI to be very useful in doing research into things like religions or folk medicine or medieval life, that I want to portray, without using any of the AI response directly.
For example, once (for a story) it took me a whole day to research how linen was made from the flax plant. Now you can do that in five minutes, and get videos of the steps in the bargain.
To me that is a useful application for AI. And of course the AIs available for spelling, punctuation and grammar.
AI story plots, on the other hand, are going to be derivative by definition, and likely bland as well because they are the average of plots.
AI is not creative, it is strictly derivative of what it was trained on.
But good luck. Always be aware that the AI may include copyrighted passages in its material, especially if you think it sounds wonderful. It may very well be wonderful, and belong to somebody that can sue you -- whether you knew it was copyrighted or not.
Just wanted to add an update/addendum to this (Nov 2024): according to the copyright office of the US,
[Generative AI] technologies “train” on vast quantities of preexisting human-authored works and use inferences from that training to generate new content. Some systems operate in response to a user’s textual instruction, called a “prompt.” The resulting output may be textual, visual, or audio, and is determined by the AI based on its design and the material it has been trained on. These technologies, often described as “generative AI,” raise questions about whether the material they produce is protected by copyright, whether works consisting of both human-authored and AI- generated material may be registered, and what information should be provided to the Office by applicants seeking to register them.
These are no longer hypothetical questions, as the Office is already receiving and examining applications for registration that claim copyright in AI-generated material. For example, in 2018 the Office received an application for a visual work that the applicant described as “autonomously created by a computer algorithm running on a machine. The application was denied because, based on the applicant’s representations in the application, the examiner found that the work contained no human authorship. After a series of administrative appeals, the Office’s Review Board issued a final determination affirming that the work could not be registered because it was made “without any creative contribution from a human actor.”
More recently, the Office reviewed a registration for a work containing human-authored elements combined with AI-generated images. In February 2023, the Office concluded that a graphic novel comprised of human-authored text combined with images generated by the AI service Midjourney constituted a copyrightable work, but that the individual images themselves could not be protected by copyright.
Emphasis mine. What this verdict seems to suggest is that the actual content of the graphic novel - i.e. the text and story, which could be represented by many different possible images and visual depictions - was copyrightable, but the specific images and depictions used - generated and, again, only one of many such images or depictions that could be used to describe the same story - were not copyrightable.
In other words (and take these with a good heaping of salt, since I don’t work for the US copyright office):
If I write a novel with the help of generative tools, for example by sticking some character descriptions and a situation in and seeing what comes out as inspiration, I think that would likely be copyrightable because I was a human actor that produced the story. I was still the key actor in writing the story; it would not exist if I did not participate. Note that my methods in this scenario also avoid plagiarizing existing copyrights because I’m not copying what the AI says verbatim, I’m just taking the idea of what it produces into consideration.
If I generate some art for the novel because I’m a bad artist and don’t want to learn how to draw crazy aliens or impossible magic effects, then I cannot copyright the art because I didn’t actually do any of the work of producing it, but I can still copyright the story because I still wrote that myself. I was not a key actor in making the art; I prompted the generator but I did not actually create the art myself.
If I generate a novel with minimal tweaks on something like NovelAI, then I probably wouldn’t be able to copyright that because it was mostly just the AI doing its own thing with my prompting rather than me writing the story. I was not a key actor in writing the story; anyone else could’ve put in similar prompts and gotten similar outputs, and I didn’t actually write anything except the prompts.
There seems to be a pattern that’s pretty easily discernible. I hope this helps.
