Who Owns Your ChatGPT content? A Dive into the Terms of Service

Companies are scrambling to figure out how to take advantage of ChatGPT and the potential it offers. But, as with any content, there are key copyright questions and risks.

Do you own the content that ChatGPT generates? While the public debate centers largely on whether AI will better the world or lead to Skynet, geeky technology lawyers like me tend to focus on the minutiae first. After all, what’s cool, revolutionary new tech without lawyers? (Don’t answer that.)

I read OpenAI’s current Terms of Use (TOU) and other policies. Since I’ve drafted many TOUs over almost 30 years of practice—and litigated a lot of agreements—a few issues stood out. For this post though, I’ll just focus on output ownership.

Do you own the ChatGPT output? Yes. Kind of. Well, it’s complicated. TOU Section 3 addresses ownership:

“You may provide input to the Services (‘Input’), and receive output generated and returned by the Services based on the Input (‘Output’). Input and Output are collectively ‘Content.’ As between the parties and to the extent permitted by applicable law, you own all Input. Subject to your compliance with these Terms, OpenAI hereby assigns to you all its right, title and interest in and to Output. This means you can use Content for any purpose, including commercial purposes such as sale or publication, if you comply with these Terms.”

It's all Conditional

This language is trickier than it looks. OpenAI uses those magic legal words assigning “all its right, title and interest” in the output to you “for any purpose,” including commercial usage. Great, right? Except there’s that pesky conditional language at the end: “If you comply with these terms.” There are many terms, but I’ll focus on just one to show how it can get complicated: Output attribution.

Section 2(c) contains eight restrictions, one of which reads: “You may not… represent that output from the Services was human-generated when it is not or otherwise violate our Usage Policies.” This means you can’t misrepresent that you created the output when ChatGPT actually did. Ownership is not the same as attribution. It can be if you personally create the content, but not always. For example, while Disney now owns Star Wars, George Lucas created it. If you try to take credit here, you would violate the TOU and void your ownership.

The TOU also doesn’t state that it must be an express representation. Implied ones may suffice. For instance, if you file to copyright ChatGPT output at the U.S. Copyright Office and claim authorship yourself, it would be an express misrepresentation and violate the TOU. (Whether the output is even registrable is an issue that the Copyright Office is addressing currently—but it’s probably not.)

What if you used some ChatGPT output on your company’s website? Your name may not appear on it per se, but if it’s incorporated with other content that you personally wrote, it could constitute an implied misrepresentation and void ownership. But then again, maybe not. If ChatGPT merely awakened your inner muse and you made significant changes to the output, ownership may still vest in you. Ambiguity abounds, and this is just a small and limited example 

Does this mean you should attribute the output to ChatGPT? That’s the obvious implication…until you get to Section 9(b), that is: “You may not use OpenAI’s or any of its affiliates’ names, logos, or trademarks, without our prior written consent.” The company’s permission appears to be a prerequisite for proper attribution. What happens if you don’t notify it? Is that another violation of the TOU? Or what if you inform OpenAI and it doesn’t consent? Is it still a misrepresentation if the company won’t let you apprise the public? And how does that affect ownership? Plenty of writers so far have freely attributed at least some of their content to ChatGPT—if for no other reason than to showcase the capabilities of this cool new technology. Notwithstanding this type of attributive descriptive use, at what point will OpenAI try to enforce this TOU provision? How selective will it be? And if a violation is found, will ownership of that output be at risk? There are far more questions than answers at this stage.

ChatGPT’s growing use and acceptance will raise all sorts of interesting issues, especially as the technology continues to improve. Ownership is only one term in the TOU, but there are many others that will test boundaries too. Like it or not, we’re all along for the ride. And before you ask: Yes, I wrote this post myself without any help from ChatGPT. Isn’t that right, Alexa?

Questions about intellectual property rights? Looking to protect your business? Reach out now to set up an appointment.