The use of Artificial Intelligence tools in academia raises two copyright-related questions:
The U.S. Copyright Office currently examines the copyright law and policy issues raised by artificial intelligence technologies, including the use of copyrighted material to train AI from the internet without permission from creators. Some copyright holders have sued the makers of AI tools for alleged infringements.
Fair use doctrine: Fair use is an important part of these debates. Researchers often rely on "non-generative" AI technologies to analyze large datasets, including texts, images, film, music. In order to be able to do so the Fair Use clause should be applied.
UC Berkeley Libraries' response to the U.S. Copyright Office’s Notice of Inquiry regarding artificial intelligence and copyright.
Fair Use could justify the training of AI tools with in-copyright material but contractual restrictions for licensed material may restrict such use. The UC Libraries in their license negotiations with publishers and other resource providers try to have use for Text and Data Mining (TDM) as well as to train AI included in licenses. However, when using library licensed material to train AI tools it is suggested to consult the Terms and Conditions of the respective license before starting.
In March 2023, "Copyright Registration Guidance: Works Containing Material Generated by Artificial Intelligence" has been released. It states that "copyright can protect only material that is the product of human creativity. Most fundamentally, the term “author,” which is used in both the Constitution and the Copyright Act, excludes non-humans."
But technological tools, including AI, can still be employed if the human had control over the work's expression. In such cases, copyright will only protect the human-authored aspects of the work. For example: Images created with AI are not copyrightable, but the arrangement, selection, or editing by a person may be.