Copyright Basics: Copyright and AI

Copyright protection is only offered to works with human authorship. This creates issues when works utilize elements generated by artificial intelligence technologies. In 2024, the U.S. Copyright Office began publishing a multi-part report on copyright and artificial intelligence that examine the copyright law and policy issues raised by AI.

In Part 2: Copyrightability, the Office concluded that existing doctrines are “adequate and appropriate to resolve questions of copyrightability” (U.S. Copyright Office, 2025, page 41). Sufficient human input for a work may warrant copyright protection, however, the Office states that prompts alone are unlikely to satisfy this requirement. More information can be found on the U.S. Copyright Office’s webpage for Copyright and Artificial Intelligence.

Due to the lack of decisions recognizing copyright protection for AI-generated works, there is no clear rule for who might be identified as the true “author” of such a work. This makes it incredibly difficult to determine who would own the copyright for an AI-generated work (in the case that protection is warranted).

Some companies, like OpenAI (creators of Chat GPT) attempt to assign copyright to the user via the Terms of Use. A company might likewise use a similar method to retain such rights.

The training of AI systems involves copying and exposing the program to mass amounts of data, which may include copyright protected works. There are also concerns that these programs may be used to generate works in the style of original creators of the works used for training, and the potential copyright implications for these scenarios.

There is still no definitive answer about whether the works generated by AI are considered infringements of copyright, as decisions are made on a case-by-case basis. As courts see more cases involving AI, we may gain a better understanding of the trends within judicial opinions.

Nothing in this guide or the linked resources is intended to provide legal advice.