Catholic University Law Review


Generative AI has created much excitement over its potential to create new works of authorship in the literary and graphical realms. Its underling machine-learning technology works by analyzing the relations among elements of preexisting material in enormous databases assembled from publicly available and licensed sources. Its algorithms “learn” to predict “what comes next” in different types of expression. A complete system thus can become glib in creating new factual summaries, essays, fictional stories and images.

A number of authors of the raw material used by Generative AI engines claim that the machine learning process infringes their copyrights. Careful evaluation of actual and likely claims shows that such plaintiffs claiming infringement will have a hard time proving reproduction, distribution, display, or preparation of derivative works and thus are unlikely to be able to establish copyright infringement under established doctrines. Fears of uncompensated appropriation, however, are likely to fuel the erection of more pay walls around original content and more licensing collectives.