•  
  •  
 

Journal of Intellectual Property Law

Abstract

AI-generated music has garnered significant attention, with many raising concerns about the new technology’s potential impact on both the music industry and human creativity. Others, including some musicians, have expressed excitement about AI and its potential for furthering creativity. Much focus has been placed on “deepfakes” of famous musicians, but there is also the issue of the use of copyrighted songs as training inputs for AI music generators. Some have argued that this machine learning process constitutes widespread copyright infringement, though it is not clear how courts will address this novel phenomenon. This Note examines AI-generated music through the lens of traditional copyright doctrines, analyzing the viability of fair use and de minimis defenses to infringement claims against AI-generated music. Special concerns relating to the Ninth Circuit’s concept of intermediate copying and AI-generated music will also be considered. By analogizing generative AI to digital sampling and considering the basic goals of copyright law, this Note will argue that there are policy reasons for findings of no infringement where the final output of the AI is not substantially similar to any of its inputs.

Download

Share

COinS