
Water Benjamin – a renowned German cultural critic – was concerned with the effects of technological advancements on human perception and culture as early as the 1930s in The Work of Art in the Age of Mechanical Reproduction. Ninety years after, this discussion has a renewed relevance as the advances of Artificial intelligence (AI) in reproducing new texts and pictures are present on a never seen scale and quality.
Author: Harri Heikkilä
Benjamin (2008) claimed that new technology – allowing mass reproduction of art – decreases the “aura” of art. Aura derives from the unique experience of an original work it offers to the viewer. Benjamin criticised the myth and rituality involved in the process and saw new ways of reproducing art as a democratising force that removes barriers to access. He claimed that technology makes art more widely available and accessible to the masses and allows art to become a form of cultural expression for all. He was also interested in the transformative nature of art and its potential to challenge traditional notions of permanence and persistence of “fixed” auratic art. Benjamin saw this development even as an emancipatory path where consumers could become producers themselves.
Killing the aura, burying the permanence
It is evident that in the era of the internet, the processes discussed by Benjamin reached a new level. The idea of Gutenbergian parenthesis by Danish professor Ole Sauerberg takes a step further, seeing the time of “fixed art” as quintessential to the Gutenbergian era. For Sauerberg, the written art in the era after Gutenberg is on a constant move, building on sampling, re-contextualising, borrowing and reshaping. (Pettitt 2007.) Interestingly Sauerberg sees here a return to the ethos of the pre-Gutenbergian time of oral tradition, where all artworks were collective, unstable and in constant change.
Figure 1. Sauerberg sees the strict authorship and stability of written art belonging only to the Gutenbergian period (Pettitt 2007, 2, modified by Harri Heikkilä)
In my opinion, the regenerative art of the 2020s’ which uses existing material to create new artworks, brings Sauerbergs’s remarks of Post-Gutenbergian era into whole new plateau: AI endlessly resamples, remixes, reborrows, reshapes, and re-contextualises. It effectively buries the auratic art and opens creative possibilities for people who have remained consumers of visual art until this time. A common man can create new art just by writing instructions – prompts – to AI. Benjamin would have rejoiced.
But artificial intelligence doesn’t merely calculate averages and probabilities from a nameless digital universe the create new; it uses the works and style of existing authors to do this. And this will pose a problem. To make my point more concrete, I made an AI example to demonstrate this by recreating “new” photographs for a known artist. Robert Frank was a photographer of the 50s beat generation who had a very distinctive style, composition, choice of subjects and mastered the greyscale sovereignly. His famous The Americans -book (Frank 1986) is an easily identifiable masterpiece.
Picture 1. Four “photographs” were created in Midjourney in May 2023 with the prompt “Robert Frank style, black & white, 1200 ASA, the Americans” (Images: Harri Heikkilä, created with the Midjourney Bot)
I asked AI to create black & white photographs of Americans in the style of Robert Frank. The result can be seen in Picture 1. You must really know your Frank to claim that these pictures never existed. And who will distinguish these “fake Franks” from genuine ones after 25 years? Does the EU norm of citing cover this kind of use? Is this what is meant by “pastiche” in the new IPR law?
But who is speaking?
There is no question about it: AI will move us outside the Gutenbergian parenthesis, but will it also bury the authorship and originality with the aura of art? And what it means if it does? Harari (2023) goes as far as seeing AI as a hacking system of the whole civilization since it gives the essence of humanity to machines: creating new texts and pictures.
Art is a commentary on and of life. If machines will comment and reshape more and more text and visual data in the future, it will narrow the authenticity because machines will end up commenting themselves. Authorship is essential in art because art is a human’s comment on life.
Technology cannot make it non-existent, but we can, at the very least, change our attitudes toward it. Technology is not fixed; it can be reshaped and regulated. At least an image made by artificial intelligence should be marked as completed by artificial intelligence in, for example, student work. Transparency and jointly agreed operating principles are necessary. One of the main principles should be the ability to opt-out from AI. It has been suggested that one could include such an option “NOAI” in metadata, but it is unclear how effective this will be in practice. The Finnish law for IPR changed in the spring of 2023, allowing now “parody” and “pastiche” use of published artwork. Also, in this case, the real-world boundaries will be set in the following years of practice.
References
Benjamin, W. 2008. The Work of Art in the Age of Mechanical Reproduction. London, UK: Penguin Books.
Frank, R. 1986. The Americans. New York, USA: Pantheon Books.
Harari, N. 2023. Yuval Noah Harari argues that AI has hacked the operating system of human civilisation. The Economist. Cited 12 Jun 2023. Available at https://www.economist.com/by-invitation/2023/04/28/yuval-noah-harari-argues-that-ai-has-hacked-the-operating-system-of-human-civilisation
Pettitt, T. 2007. Before the Gutenberg Parenthesis: Elizabethan-American Compatibilities. Cited 9 Jun 2023. Available at https://www.academia.edu/2946207/Before_the_Gutenberg_Parenthesis_Elizabethan_American_Compatibilities
Author
Ph.D. Harri Heikkilä is a principal lecturer of visual communication at the LAB University of Applied Arts.
Illustration: Photo from the Jaume Plensa exhibition in 2015, Le Musée d’Art Moderne de Céret (Photo: Harri Heikkilä)
Published 26.6.2023
Reference to this article
Heikkilä, H. 2023. AI in the Age of Reproduction. LAB Pro. Cited and date of citation. Available at https://www.labopen.fi/en/lab-pro/ai-in-the-age-of-reproduction/