Prof Annemarie Bridy Asks: How Human Does An Author Need To Be?

Brent Randall is a JD candidate at the University of Ottawa.

The Canadian Copyright Act does not explicitly define the term “author”, but the statute does appear to assume that the “author” will be a human. A recent paper by Professor Annemarie Bridy seeks to challenge this assumption as she argues that “all creativity is inherently algorithmic” and therefore works produced by computers should not be viewed as significantly different from those produced by humans.

In her paper, Coding Creativity: Copyright and the Artificially Intelligent Author, Professor Bridy is not focusing on the creative ways humans interact with computers to author works. Instead, she concerns herself with “the ways in which people are enabling computers to produce art and other creative works in new ways, virtually all by themselves.”

In the 1884 case, Burrow-Giles Lithographic Co. v. Sarony, the Supreme Court of the United States considered the extension of copyright protection to photographs, despite the creation of such works relying heavily on the camera and not just the human.  The Court refused a judgment that would make the camera the author in such instances and instead used a causation argument to find that the author is “the person who effectively is as near as he can be, the cause of the picture which is produced…”

Building on Burrow-Giles was the 1903 Supreme Court of the United States decision, Bleistein v. Donaldson Litographing Co., in which it was affirmed that, “[t]he copy is the personal reaction of an individual upon nature.  Personality always contains something unique.  It expresses…something irreducible, which is one man’s alone. That something he may copyright…”

The cases of Burrow-Giles and Bleistein, Professor Bridy argues, “marks the jurisprudential moment at which copyright protection became virtually guaranteed for any work produced by human hand, regardless of its perceived creativity or aesthetic merit.”   But what, exactly, creativity is has not been clearly explained with the closest attempt coming in the 1991 United States Supreme Court decision, Feist Publications, Inc. v. Rural Telephone Service Co., Inc.

Professor Bridy focuses heavily on trying to figure out the definition of creativity and whether artificial intelligence (AI) is capable of being creative or, if by its very definition, creativity implies the work of a human.  If, for the sake of argument, creativity does not require a human, what would a computer have to do to produce something creative and therefore potentially be considered an author? Must it produce results that are unpredictable? Can a computer do anything unpredictable, if it must first be created and programmed by a human, or is that just a more complicated instance of human causation as seen in Burrow-Giles? What follows for Professor Bridy is a deeply philosophical exploration of the meaning of thought and the workings of the human mind.  Cognitive scientist Marvin Minsky, for instance, has argued that the brain is nothing more than a “meat machine”, presumably subjected to its own kinds of “programming” from parents, society, and education.

The ultimate conclusion for Professor Bridy is that the American “work made for hire” doctrine could be applied in cases in which an author is a computer.  The United States Copyright Act of 1976 defines work for hire quite similarly to Canada’s Copyright Act’s section 13(3), which states that, “[w]here the author of the work was in the employment of some other person…the person by whom the author was employed shall…be the first owner of the copyright.”  Looking at creation in this way avoids complications with deciding who – or what – the author is, since it essentially “bypasses” their rights associated with the work (except for moral rights, which likely would not matter to a computer, at least at this point in the development of AI).

What one’s definition of “creativity” includes ultimately has a major impact on the definition of authorship.  Such a definition would likely involve concepts of unpredictability and novelty.  But even this definition causes problems for AI authorship, at least for some, since there is a question of how unpredictable and novel creations can be when they are made by a machine built and programmed by a human.  However, it seems that computers are capable of behaving unpredictably (we see this in video games, for example) and as such, their creativity may be enough for authorship.  Regardless of one’s definition of creativity and authorship, it seems that the status of works created by AI may not matter as much if the copyright in such works is ultimately owned by human employers.

One Comment
  1. Under the US “work for hire” doctrine, a corporation may in fact be an author, but that is not true under the Canadian Copyright Act, where an employer may become only an “owner” of copyright, never an “author” (otherwise the human author loses all moral rights). This is not a matter of assumption, but decision: Massie & Renwick Ltd. v. Underwriters’ Survey Bureau Ltd. [1940] S.C.R. 218. There is a vestigial exception in s. 10 of the Canadian Act, which allows a corporation to be the author of a photograph, going back to the the 1921 Act, where sound recording makers and photographers were “deemed” authors because of their ownership of the physical material on which their work was fixed. This vestige will likely be repealed in the next round of copyright reform. Even under the US Act, the corporate author gets that status as a matter of derivative title from its employee or other specified freelancer – so one must ask who (not what) that entity was. The entity must be specified or inferred.
    The question whether machines, or animals for that matter, can be authors is less a matter of trying to squeeze them into the current law, whether American or Canadian, which is quite unaccommodating, but rather whether the statute should be amended to include them. The humanistic case is here pitted against the scientific. Science partly won out in the UK, which introduced a category of computer-generated (un-authored) works in its Copyright Designs & Patents Act 1988. One must provide for the duration of their copyright (which cannot be predicated on “life” of machine plus a number of years) and also of which state (if any) it is a national (for Berne and TRIPs purposes) – presumably not the place of its manufacture; and for Canada, whether moral rights may be claimed.
    I doubt that a court can legitimately juggle computers or animals into the current law by a process of “interpretation”, although A.P. Herbert once had his fictitious plaintiff Haddock sue a computer in one of the Misleading Cases: Haddock v. The Generous Bank Ltd, Computer 1578/32/W1, the Magical Electronic Contrivances Ltd, and the Central Electricity Board: Reign of Error? (1963). Some years ago too, a US court decided that Blackie the Talking Cat was not a “person” whose free speech rights were protected by the First Amendment to the US constitution: Miles v. City Council of Augusta, Georgia, 710 F.2d 1542 (11th Cir. 1983). Perhaps it might hold differently today for Lenovo the Thinking Computer.

Comments are closed.