•  
  •  
 

Abstract

U.S. IP law has grappled with the issue of non-human authorship and inventorship since the middle of the twentieth century when computer systems were first used in the generation of music and imaginative ideas. Today’s advanced Artificial Intelligence (AI) systems have “created” numerous works including musical compositions, art, writings, recipes, and potentially patentable inventions. However, common-law, along with the policies and procedures of the U.S. Copyright Office and the U.S. Patent and Trademark Office, rejects the idea of non-human authorship or inventorship. These doctrines are not based off statutory requirements but on assumptions about computer capabilities stemming from an analysis done in the midtwentieth century, almost 40 years ago. Other contributors to this question have denied that AI authorship should ever be allowed or that AI should be treated as inert tools of creation no different than cameras or photocopiers. However, both approaches fail to appreciate the independent creation of modern AI’s that process information in ways much like human brains—well beyond simple mechanical devices using simple mathematical algorithms. Thus, this paper argues for recognition of AI authorship and inventorship. To that end, this paper presents a framework to analyze when such rights should be recognized based on (1) the subject matter (i.e., independence of the work) and (2) the causation (i.e., independence of the AI). In addition, the framework introduces an IP rights assignment regime like “work-for-hire” and “employed-to-invent” based on the nexus between the AI and the natural persons programming and/or using it. This approach avoids the complication of naming an AI as a legal or natural person under the law.

Last Page

38

Download

Share

COinS