Fluxus and AI, An Interesting Mix

 

Take a moment and think about the most Fluxus artwork you can. Did you imagine Yoko Ono’s Cut Piece? Or perhaps John Cage’s 4’33? What I’m not going to do here is tell you that I’ve found an artwork that exemplifies the Fluxus ideals more than those pieces, though I am tempted to do so. But, what I am going to do is share a work that I find uniquely Fluxus and allow you to compare it to your vision of Fluxus.

Fluxus.AGES is an artwork created by Émilie Fortin and Pat McMaster, a pair of Montreal based interdisciplinary artists, in 2020. The work itself is a web page and is easy enough to find through Google or going to the URL fluxus.jagra.ca. The page that greets the viewer upon first visiting the site briefly summarizes the work as a “collection of performance event score experiments and artwork created in collaboration with artificial intelligence.” Also featured on this page is the spelling out of the acronym that takes part of the title of the work as being algorithmically generated event scores. But, the last, and perhaps most important, part of this page is the begin button. Upon pressing said button, a reader is brought to another page. This page features a titled event score on the right hand side and either what looks like a scanned piece of black and white film or a framed and captioned colour photo or painting. Under the event score are four labelled buttons. One to go to the previous score and one to view the next one. A button to return to the homepage and a button labelled random. There are around seventy event scores that can be seen by pressing the next button, each prompting the reader to do uniquely Fluxus acts such as licking a number of household objects and then making sure their knife is sharp. Going back to that last button though, the random button is where the true nature of the work shows its face. Pressing the button brings the user to another page with a titled event score and a captioned picture. But, even if the viewer had memorized all seventy or so scores from before, this score should be completely new to them. In fact, every person that visits the site should be able to
see a completely unique set of event scores.

Algorithmically generated event scores. The title of the work gives away a lot about the nature of the piece. Each time a user presses the random button they are greeted by a new set of event scores with titles and captioned images alongside. Each event score seen by the user was pulled from the output of three different AI models that had all been fed fifty event scores as reference material. And much like the event scores themselves, the images beside each score were also generated by a computer. Feeding each event score through a text to image generator creates the surreal images that always resemble something, but they can never be fully recognized as representational imagery. The captions for said images were produced by feeding those images through yet another algorithm before being rendered in handwriting by another AI model. And just like that, the user is greeted by yet another set of event scores. Unfortunately though, it seems like this site doesn’t do any of that processing in real time. This means that every page is pulled from a set of old data instead of being generated again for each individual user. While this does mean that there is a set number of event scores, this limitation doesn’t seem to limit the validity of the work as being uniquely Fluxus in nature.

Mass-production, ephemerality and chance aesthetics. Those three qualities of the Fluxus movement are exemplified through this work. If our current age of mass consumption and rampant consumerism has taught us anything, besides the dangers of a society built on these qualities and how much we really shouldn’t put faith in those with power, it’s that computers are far faster than humans at creating a product. While there may be a limited number of event scores archived on the website, the number of unique scores could number well into the thousands with the means to create significantly more still a possibility. And what, I ask you, can truly be more ephemeral than bytes of data that never existed as more than the magnetic alignment of a metal platter or a collection of photons emitted from a screen. But chance aesthetics, as it is the more defining trait of the Fluxus movement, is where this work seemingly is in a league of it’s own. From start to finish, each stage of this artwork removes the human factor in favour of letting a computer make all the critical decisions. The computer creates the event score before creating an image out of said score and captioning that. There is true randomness present in the decision-making of a computer that humans can’t come close to. But, the human factor is far more present in this artwork than one would think. Both the input and output event scores were curated by humans. A person had to sift through the text generated by each AI and extract the usable ones from the mess of unintelligible things an AI model will spit out.

But does that take away from the work. Does the severe lessening of chance aesthetics through the heavy-handed curation of what is read and said by a computer kill the chance aesthetics of this work. Or is it still a work that can be called uniquely Fluxus. I won’t decide for you.

 

 

 

Works Cited

Friedman, Ken, and Owen Smith. “Fluxus and Legacy.” Visible Language, vol. 39, no. 3, 2005, pp. 212–217.

Miles, James, and Stephanie Springgay. “The Indeterminate Influence of Fluxus on Contemporary Curriculum and Pedagogy.” International Journal of Qualitative Studies in Education, vol. 33, no. 10, 2019, pp. 1007–1021., doi:10.1080/09518398.2019.1697469.

Salvaggio, Eryk. About. www.fluxusexmachina.com/about-2.