Abstract
We are on the cusp of a very significant change. Programs can now generate music that is commercially viable as background music, or write poetry that is difficult to distinguish from that written by humans.7 They can write simple news stories, such as breaking news about earthquakes or sports.8 They can generate personalized reports from databases, which is essentially what a Google search result is. Or they can create automated videos, as Facebook has done for its users occasionally.9 And they are doing so for financial gain.
What is perplexing about these situations is that they seem to be instances in which we have something that looks like a “work,” but there may be no person whose actions resemble those of a traditional “author.” That is, computergenerated works pose a problem for what might be called the “standard model” of copyright law, under which a person, the author, produces a work that is then conveyed to the audience through some sort of medium such as a book. With computer-generated works, the production and conveyance steps become intertwined, with one work—the computer program—producing another work each time it is used. Those subsequent works may contain creative elements that were not present in the first, and thus lack an easily identifiable human origin. When it comes to what might be “emergent works”—works that consist largely of creative elements that have emerged unbidden from the operation of the program—who should be considered the author?