« Articles / Who Owns ChatGPT’s Words?

Who Owns ChatGPT’s Words?

By Rushil Yarrabolu
July 7, 2025

As AI tools like ChatGPT and Gemini redefine how we write, compose, and code, one big question has still not been answered: Who owns the product? Whether it's viral poems, business plans, or courtroom briefs, the AI copy is everywhere. But in a technical sense, none of it belongs to anyone. And that poses a problem. 

 

To understand the scope of this gray area, we need to examine how copyright works—and why U.S. law just isn't prepared for artificial intelligence. 

 

No Human, No Copyright 

 

The U.S. Copyright Office explains that copyright only applies to work created by human beings. That makes sense—until you think about how much of today's material is being created, written, or edited by AI. 

In 2023, the D.C. District Court ruled that a machine-created work of art produced by a device known as "Creativity Machine" is not copyrightable, although a human being claimed ownership of the device on which the artwork was created. The decision made it starkly obvious: if a human does not write it, paint it, or make it, it is not protected. 

 

Which is to say that if you compose a short story with ChatGPT or paint an album cover with Midjourney, no one—least of all you—owns the resulting work. And since property can't be held by the AI models per se, their work is essentially in a state of legal limbo. They're automatically in the public domain. 

 

Why That's a Big Deal 

 

The legal ambiguity has profound implications across industries. For instance, if one startup writes a pitch deck using an AI tool and another startup copies it word-for-word, there is no grounds for a lawsuit for copyright infringement. If a self-published author uses AI to write a book and it becomes a bestseller, there is no legal recourse to stop it from being copied.

Worse still, if the news that a publishing source releases is produced through AI without human editing, it can be copied or printed out at will. In an era when machines can compose books, essays, and even movie scripts, the idea that none of it can be owned in the traditional sense is an enormous disruption to how we think about intellectual property. 

 

So What About Co-Creation? 

Others contend that if a human edits or heavily influences the output of the AI, then it is subject to copyright. In March 2024, the U.S. Copyright Office revised its policy to make clear that only "human-authored" parts of work assisted by AI are eligible for protection. For instance, if an artist creates a visual idea with DALL·E and then heavily edits it in Photoshop, the revised version could be copyrighted. 

 

But here's where it gets complicated. How much editing is enough? If the student has ChatGPT compose an essay and they adjust a couple of sentences, is that essay theirs? If the reporter gets an AI-generated draft of an article and tweaks the organization, does it constitute copyrighted material? 

 

So far, the Copyright Office has strictly enforced. In a 2023 denial letter to author Kris Kashtanova, it ruled that their graphic novel—mostly produced with Midjourney—was partially protected. The images created by AI? Not covered. 

 

Outside of Copyright: Liability and Ownership 

 

There is another issue as well: if no one owns AI-generated content, who's responsible for it? 

 

Let's assume that a chatbot generates defamatory, libelous, or plagiarized material. Can the user be held accountable? Can the company that owns the AI model be held accountable? The courts are in an ambivalent state. Some believe that AI software should be considered more like publishers—accountable for misinformation, hate speech, or defamatory content they help generate. Others believe the user should be held fully accountable, not the platform. 

 

Meanwhile, the big guns like OpenAI and Google are trying to have their cake and eat it too. They're offering enterprise-level indemnity agreements that promise to protect users from particular legal actions. But those protections tend to be narrow, and they avoid the very basic question: if AI is writing it, who's actually in charge? 

 

The Future of AI IP

 

There is broad momentum to revisit the regulation. Lawyers, technology companies, and lawmakers alike have proposed fresh models for dealing with AI-created content—anything from creating a new category of IP law to requiring AI disclosures on published works. Some even imagine a "copyright credit" system, in which AI companies license the training data that they use and pay out royalties thereof. 

 

Until those ideas become law, though, we’re stuck in limbo. For now, the safest path is transparency: clearly stating when AI is used, editing its output heavily, and avoiding full reliance on it for copyrighted or commercial work. 

Because, as powerful as ChatGPT may be, its words still don’t belong to you—or anyone else. 

 

Sources

 

[1] https://www.copyright.gov/ai/ 

[2] https://www.reuters.com/legal/government/us-court-says-artificial-intelligence-cannot-hold-copyri ght-2023-08-21/ 

[3] https://www.nytimes.com/2024/03/16/technology/ai-copyright-guidelines.html [4]https://www.theverge.com/2023/11/7/23948617/midjourney-ai-artwork-copyright-office-kris-ka shtanova 

[5]https://www.forbes.com/sites/maryjuetten/2024/01/28/who-owns-ai-generated-content-legal-p erspectives-and-business-risks/ 

[6] https://www.washingtonpost.com/technology/2024/06/09/openai-microsoft-legal-liability/

[7] Image: https://emaggiori.com/chatgpt-vocabulary/

An unhandled error has occurred. Reload 🗙