IP Tech Knowledgy

Works created by artificial intelligence (A.I.) raise two core copyright questions: is the work an infringement, and is the work protectable? Some answers are easier than others, and what A.I. itself writes on the subject is sometimes inconsistent.  

Increasingly, people are using A.I. robots and software as their author and artist. The ChatGPT robot is currently in vogue as a supplement — or replacement — to providing all types of written work, from articles to pleadings to term papers.

How copyright law deals with the A.I. creation is not always a clear answer, and the answers ChatGPT itself provides shows how it can be alluring, yet sometimes unreliable, in its pronouncements.

As to whether the use of material provided by a chatbot is infringing someone’s copyright, ChatGPT correctly writes: “There is no definitive answer to this question as it depends on individual circumstances. Generally speaking, using a chatbot service that scrapes data from other websites in order to populate its own chatbot conversations may be considered copyright infringement.” Answers to various prompts to ChatGPT identify correct defenses that concepts of “fair use” and using public domain materials are not infringements.  

A person who uses narrative provided by A.I. should be wary of infringing another’s copyright, because, as one ChatGPT response to a prompt state “[I]t is important to note that ChatGPT is trained on a vast amount of text data from the internet and other sources. As such, the model may generate text that is similar or identical to existing copyrighted works. It is the responsibility of the user of the model to ensure that the use of the model does not infringe on any existing copyrights.” Also, the output could be an unlicensed derivative work from someone else’s original work.  

Consider that the resources used by a chatbot are not too different from those available to a search engine; the A.I. robot output is merely synthesized better. Getting the same information from a search engine at least makes it clear that the results and information are taken from a specific source, rather than seeming like original authorship. ChatGPT also warns: “[I]t is important to check the terms of service and any agreements made with the entity who controls the use of the model to determine the specific ownership of any works created by the model.”  

The harder question is who, if anyone, owns the copyright in the work that A.I. creates. The United States Copyright law currently requires a human to have created the work. ChatGPT properly alludes to a well-known case when it responds to a prompt, writing: “A monkey cannot own a copyright.” (See the losing plaintiff HERE) The A.I. also properly responds: “ChatGPT cannot own the copyright to what it created because it is not a person or legal entity.”

There are nuances to ownership which courts are grappling with, and which other ChatGPT responses allude to, such as: “The copyright to works created by artificial intelligence is owned by the person who created the AI, unless they specify otherwise,” and “A computer can't own a copyright, but the creator of a work can.”  These responses are correct that if the programmer had enough input into the workings of the robot and the output could be tracked to the programmer’s input, then the programmer might own the copyright in the robot’s output. Otherwise, the new creation, produced by a machine, would be in the public domain and be available for anyone to use — unless it is infringing on the rights of the author from whom the A.I. scraped the work.

A.I. can also be wrong.  Prompts to ChatGPT also elicited the following responses: “ChatGPT owns the copyright in the work created by him” and “ChatGPT is a copyrighted work. The copyright owner is ChatGPT.”  

Anyone using A.I.-produced works should think about how copyright law might restrict the right to use or own the output. Congress may ultimately need to address how works by a non-human Artificial Intelligence robot can be used and owned.

Ned T. Himmelrich
410-576-4171 • nhimmelrich@gfrlaw.com