Copyright and ChatGPT: Who owns AI-generated content?
We are living in a new era of creative collaboration where asking ChatGPT to create content for us is the new normal. Using the popular Large Language Model to create blog articles for marketing and SEO purposes is an obvious use case, but we’re not stopping there. There are already hundreds of books you can buy on Amazon that list ChatGPT as the author or co-author. And probably many more are hiding the fact that generative AI played a role in their creation.
I’m not judging the humans behind this movement - I’m one of them! The Nexus Fallacy (chapter 1 and chapter 2 available online) is my own attempt at co-authoring a novel with the AI. The line between human and AI-generated content is blurring and new questions about copyright ownership need to be answered.
Everyone creating content with ChatGPT will ask themselves these questions sooner or later: Who owns the output of ChatGPT? Is it the human behind the prompt, OpenAI, or no one? And does that give us the rights to distribute and sell it?
As I dive deeper into my creative journey with ChatGPT-4 and MidJourney 5 (The Nexus Fallacy graphic novel on the way?…) , I’m learning about copyright in the age of generative AI.
Understanding copyright law
Before delving into the specifics of AI-assisted book writing, let’s look at the basics of copyright law. In most countries, copyright protection is granted to original works of authorship that are fixed in a tangible medium of expression. This can include literary works, musical compositions, visual art, and more. Copyright grants the author the exclusive right to reproduce, distribute, publicly perform or display, and create derivative works based on their original creation.
In the realm of literature, copyright protection extends to novels, short stories, poems, essays, and even e-books and digital content published online. To better understand how copyright applies to these formats, let’s delve deeper into some examples.
Print books and traditional publishing
In traditional publishing, authors typically hold the copyright to their work. However, they often grant publishers the exclusive right to reproduce and distribute the work through a licensing agreement or transfer of copyright.
Example: J. K. Rowling retains the copyright to the Harry Potter books, while the publisher has been granted specific rights to reproduce and distribute the books.
E-books and digital publishing
E-books and digital content are also eligible for copyright protection, as they are considered fixed in a tangible medium of expression. Copyright ownership follows the same principles as traditional publishing, with the author keeping their rights unless transferred to a publisher or another party through an agreement.
Example: An author who self-publishes an e-book on Amazon Kindle retains the copyright to their work, granting Amazon a license to distribute the e-book through their platform.
Online content and blogging
Copyright protection extends to original content published online, including blog articles, essays, and other literary works. Authors keep copyright ownership over their content unless they explicitly transfer those rights to another party or publish it under a different license, such as a Creative Commons license.
Example: A blogger (like myself) who writes and publishes articles on their personal website (like the one you're reading) owns the copyright to their work, protecting it from unauthorized reproduction or distribution.
Fan fiction and derivative works
Fan fiction, which involves creating new stories using characters, settings, or plot elements from existing works, falls into a gray area of copyright law. While authors generally own the copyright to their original contributions, the underlying work’s copyright owner may have a claim if the fan fiction is considered a derivative work. Sometimes, the use of existing copyrighted material might be considered fair use, depending on factors such as the purpose, nature, and substantiality of the borrowed content.
Example: A fan writes a story featuring characters from Harry Potter. J. K. Rowling may have a claim against the fan fiction if it’s considered a derivative work that infringes on her copyright, unless the fan fiction falls under fair use.
AI-generated content and copyright
Content generated by artificial intelligence is becoming more sophisticated and commonplace. However, the legal landscape surrounding copyright ownership and protection still needs to catchup with the technology. In many jurisdictions, including the United States, copyright protection requires a human author. This means that AI-generated content, without human input or creative effort, might not be eligible for copyright protection on its own.
When using ChatGPT to write a book, the extent of the user’s input and involvement can vary. I’ll explore three common scenarios. In each one, the user’s creative input and the AI-generated content have different implications for copyright ownership.
Scenario A
The user may hold copyright over the original ideas, characters, or storylines they provide. The AI-generated content, being a derivative work, might not be eligible for copyright protection on its own. The user’s claim to copyright ownership would be stronger if their input is more creative and original.
Scenario B
Since the user provides detailed instructions, their contribution to the story is more substantial. The AI-generated content remains a derivative work, but the user’s claim to copyright ownership is stronger. The final work could be considered a collaborative or joint work, with the user as the primary author.
Scenario C
By providing detailed input and also editing the AI-generated content with their creative insights, the user adds another layer of originality and creativity to the work. This strengthens their claim to copyright ownership, making them likely the author of the final work.
The last scenario reflects the process I’ve been using in my collaboration with ChatGPT for The Nexus Fallacy and I expect to retain the copyright of the book. I’m documenting the outline of the story, the characters, and the structure of each chapter on a separate Notion file. I’ll feed the specific plot to ChatGPT and provide multiple rounds of feedback until I’m happy with the output. Then I slice it and put it on a different writing tool, where I edit steadily, sometimes adding new pieces from scratch, until a chapter is ready.
This process is necessary because ChatGPT-4 is still not capable of writing a full book (or even a full chapter) with one prompt. But it also makes the creative journey more fun and ends up giving me a better claim to ownership over the final work.
Tips for safeguarding your copyright
Enforcing copyright in AI-assisted works can still be challenging due to the lack of clear legal guidelines. As the legal landscape evolves, it is crucial for authors to consult with legal professionals to understand their rights and protect their intellectual property.
When using ChatGPT to write a book, consider the following tips to strengthen your claim to copyright ownership and protect your intellectual property:
- Document your creative process, including original ideas, characters, and plot elements you contribute. This documentation can serve as evidence of your creative input if a copyright dispute arises.
- Be proactive in editing and refining the AI-generated content to ensure that your creative insights are evident in the final work.
- Register your work with the appropriate copyright office, if applicable in your jurisdiction. This can provide additional legal protection and may be required to enforce your copyright in court.
- Stay informed about the laws surrounding AI-generated content, as this can help you better understand your rights and responsibilities.
- Consult with a legal professional to discuss the specifics of your situation and obtain tailored advice on protecting your copyright.
- Review the terms of service and any applicable licensing agreements when using AI tools like ChatGPT. This will help ensure that you understand the rights and obligations of both parties, including any potential claims to copyright ownership.
In case you’re wondering, ChatGPT assisted me in the creation of this article. And now it’s calling me to go back to work on the third chapter of our book about extinction by AI.
The information I provided should not be considered legal advice. As the legal landscape around AI-generated content continues to evolve, it is essential for authors to stay informed and consult with legal professionals.