[Guest post] Emergent works without foundations

Protection of emergent works, such as autonomous outputs of artificial intelligence systems, or monkey selfies supposedly lacking a human author, is impossible. Without an author, there is no expression of ideas which in turn could be original, and thus no copyrightable work. As I argued recently, this is the starting point of the whole copyright project: the conceptual building blocks of idea-expression dichotomy, originality, authorship, and the concept of a protectable work, the international legal framework, and the doctrine in both the United States and the European Union begin and end with the figure of a human creator. This is a framework which allow for a lot: but not for emergent works.

From DALL-E: “Kat dressed up as a judge wearing a wig and with angry expression”

At least up until recently, it was fashionable to deconstruct copyright theory and doctrine, claiming it is political, explicitly humanist, and incompatible with postmodern literary theory; sometimes with a neuroscientific addition that, since all humans are like algorithms, there is no difference between us and computer programs. These authors are right, to an extent: copyright does create a distinct reality, formed by its causal concepts. It is a reality which stems from a concrete historical-philosophical moment, which is not culturally or politically neutral. To protect expressive works as separate from material reality of objects, that which is original in the “modicum of creativity” or “author’s own intellectual creation,” and to advance concrete goals, is to create a system which excludes some subject-matter and some policy proposals – such as emergent works.

The recent decisions of the U.S. Copyright Office’s Review Board in A Recent Entry to Paradise, Zarya of the Dawn [IPKat here and here], and the District Court judgement in Thaler v. Perlmutter [IPKat here] all embraced traditional copyright causation and its human-centeredness, declining protection of emergent works. This was, both a doctrinal inevitability, but also a conceptual one, stemming from legal theory reinforced by the international framework of the Berne Convention, the TRIPS Agreement, and the WIPO Copyright Treaty.

To protect AI creations just because they are costly is a new version of the labouristic fallacy, or of the “if value, then right approach;” to protect them merely because of aesthetic even is similarly misguided. After all, just because a reproduction looks like the original, it does not mean its copyrightable. Copyright’s causation and all of its language makes this clear: that which is not an original, creative expression does not exist as a matter of copyright’s positive ontology. It’s just like nature, things, or ideas; can be creatively transformed, but of itself, does not entitle to anything. At the same time, it can also be infringing, like any unoriginal reproduction.

We are yet to see test cases on the European continent, quite possibly because the Court of Justice of the European Union’s recent jurisprudence has made the personal, creative expressive contribution the be-all and end-all of copyright: a requirement of originality and of the concept of work alike. Autonomous outputs of AI, or those of monkeys, robots, or other non-humans are “no properties.” The law of copyright simply does not ascribe meaning to the acts of non-humans, does not recognize their aura, and provides no monopoly rights to the law’s non-subjects. To argue otherwise is to undermine the whole of copyright’s paradigm.

Of course, in addition to the legal-theoretical, there is another dimension to the unprotectability of robotic copyright absent human original expression. Copyright, like all of modern law, promotes human welfare. It does so in the cultural domain, allowing us to live good lives, as members of the cultural community. We do not need to provide fundamental reasons for our worth, existence, or the value of human culture; copyright law does it for us.