AI, Machine Learning and Art

Should the creative outcomes/ artifacts generated by AI be considered “the same” as those by humans if they are identical down to molecular level?

I’m not exactly sure yet how to answer this question. For me personally, art is a human tradition and thus is defined by human agency. The reason why “art” exists is that during the course of history, we developed an intelligence that allows us to distinguish our actions from instinctive behavior and spend our energy on activity other than fighting for survival. For me, it is not the outcome but the intention of art that matters. I believe that creative people create art to express themselves and negotiate their place in the world and in society. I cannot see this intention in computers and I would thus be inclined to consider creative artefact created by computers as “meaningless” as they don’t fulfill the purpose of self-expression. However, I would never argue that they are of lesser quality because they are for sure pleasing to the human eye/ ear/ senses in general. I am aware, though, that there is no clear-cut boundary between human art and AI art. What about the programmers and developers of the AI? And what does this mean for computer assisted art then?

I think that Benjamin’s concept of aura can also be applied to AI and computer art although he relates it more to reproduction of artworks. I would argue that part of the “aura” of an artwork comes from the intention and personal expression behind it, which again locates it at a specific point in history. It seems to me that the biggest problems we have with the idea that art could be generated by computers instead of humans is that we have theory of mind to put ourselves into other people’s shoes and understand their feelings and motivations, but we don’t have any way of thinking like a computer and understanding why it does what it does. A computer has no intentions other than the ones it is meant to have, even if it is able to learn from and interact with its surroundings. The authenticity and uniqueness of a work of art which Benjamin attributes to its traditional context, can be replicated by a computer, but the traditional context is still a human one. Benjamin describes the “original use value” [1] of the artwork as ritual or ritual function, which has to come from an intent, that I believe computers cannot develop.

I would categorize AI generated art as a third-order simulacrum in the Baudrillardian sense:

“The third stage masks the absence of a profound reality, where the sign pretends to be a faithful copy, but it is a copy with no original. Signs and images claim to represent something real, but no representation is taking place and arbitrary images are merely suggested as things which they have no relationship to. Baudrillard calls this the “order of sorcery”, a regime of semantic algebra where all human meaning is conjured artificially to appear as a reference to the (increasingly) hermetic truth.” [2]

AI art pretends to be located in a history and tradition of human thinking without being able to be part of this history and tradition itself.

Nevertheless, reading Boden’s article [3] and learning that Cope destroyed Emmy, an AI capable of composing music of human quality, and all the complete database after working on it for 25 years due to missing appreciation of its artistic value, made me sad. One of the reasons for the destruction of Emmy was the fact that people did not regard the compositions as music but merely as computer output, which would be my way of thinking, too. I would say that this attitude comes from out generally sceptic stance towards computers and machines. The destruction of Emmy, which ironically made the AI resemble a human composer even more by giving it a finite oeuvre, made me think of the movie Bicentennial Man with Robin Williams [4], about the desire of a robot to be recognized and die as a human.


[1] Benjamin, W. (1969) ‘The Work of Art in the Age of Mechanical Reproduction’, in Arendt, H. (ed.), Zohn, H. (tran.) Illuminations. New York: Schocken Books.

[2] Anon 2019. Simulacra and Simulation. In: Wikipedia. [online] Available at: <https://en.wikipedia.org/w/index.php?title=Simulacra_and_Simulation&oldid=887096205> [Accessed 21 Mar. 2019].

[3] Boden, M.A., 2009. Computer Models of Creativity. Association for the Advancement of Artificial Intelligence, pp.23–34.

[4] Bicentennial Man. 1999. Directed by C. Columbus. Buena Vista Pictures; Columbia Tri Star.