AI art is a remix, and we are experiencing an epistemological crisis about it because we don’t have a culturally sanctioned status for the remix as art.
Digital and electronic cultures, dance music standing out as the primary example, have for decades understood art in more layered fashion than the current mainstream paradigm.
In the underground cultures of dance music, originals can be repurposed endlessly as ”building blocks” for new layers of art, in the form of the remix and the art of the DJ.
These cultures do not reject the idea that a remix or a DJ has a purpose of it’s own, or claim that only the originals matter like some mainstream art critics or record industry strongmen have recently stated. The performance: the order, the emphasis, the presentation, is what matters. The story told by the remix artist with the originals has purpose, if it speaks to the listener’s sensibilities. Bad remixes don’t excite anyone, these are called ”train wrecks” for a reason.
The skill and the art of the DJ as an artist in their own right however, has not been properly acknowledged by the mainstream art paradigm. Even if the reality on the ground is that most art we consume is indeed a remix of some kind, the concept of ”original” has to be shoehorned on everything nevertheless. You are either a copy or an original, there is no in-between.
Enter AI art tools, and the moral panic surrounding them: Effectively the machine learning tools do nothing dissimilar to what the DJ or remix artist were already doing. They allow the user to become a ”DJ of pictures”, to tell a new story by using existing pieces of art as tools.
The moral panic is largely an epistemological crisis: We don’t have a socially acceptable status for the legibility of the remix as art-in-it’s-own-right. Instead of properly appreciating the remix and the art of the DJ, the remix, or the meme cultures, we have shoehorned all the cultural properties associated onto an 1800’s sheet music publishing -based model of artistic credibility. The fit was never really good, but no-one really cared because the scenes were small, underground and their breaking the rules was largely out-of-sight. In the case of Hip-Hop music, the issues of licensing beats were pushed into the background, while the rapper took the mantelpiece of ”the original artist”. Controversies with sampling were discussed as anomalies, from which culture largely rubber-banded back into the old model.
However, the future is not bleak: The remix, the DJ-as-artist, bootlegging and other relevant cultural phenomenons were widely theorised in journalism and socio-/musicological academia of the 1990’s and early 2000’s.
As the AI art tools enable the masses to become in the pictorial and written arts as what DJ’s did to the music world, we should look in to the writing some distinguished scholars first before rushing in head-long to paint a picture of what’s happening in the terms defined by whatever the current media zeitgeist supposes us is important.
Recommended reading:
- Graham St. John
- Erik Davis
- Anthony D’Andrea
- Simon Reynolds
- Bill Brewster & Frank Broughton
- Kai Fikentscher
- Walter Benjamin’s classic 1935 text: The work of art in it’s age of mechanical reproduction (here the reader should remember read it in it’s pre- World War II context)
I concur that the AI art tools are simply resurfacing an old problem we left behind unresolved during the 1980’s to early 2000’s. Now it’s time for us to blow the dust off these old books and apply what was learned to the situation we have at our hands now.
We should not forget the modern electronic dance music industry has already developed models that promote new artists via remixes of their work from more established artists. These real-world examples combined with the theoretical frameworks above should help us to explore a refreshed model of artistic credibility, where value is assigned to both the original artists and the authors of remixers, who use their originals to tell a new story, fitting the particular life-story of the particular viewer. Like a deejay spins just the tracks you needed to hear at that particular night of your life at that particularly important moment, the value of the experience encapsulates both the original artform and it’s application to the particular context.
To fully appreciate and integrate AI art in our culture, we cannot rely only on our established models of artistry and credibility. From what was once only a fringe endeavor of collague or plunderphonics artists, mass production tools have forged a mainstream phenomenon. This is however, not our first contact with art like this, and we do have the theoretical frameworks available to form a new class of art, if we reach just a little further into the long corridors of university libraries and humanities departments for them.
For a more technical & legislative approach to the issue, please see my previous blog post: https://gimulnaut.wordpress.com/2023/01/13/copyright-wars-pt-2-ai-vs-the-public/
————————
Afterthought 1: The transformation function in art
Art, especially popular forms of it, has always been a lot about transformation: Taking what exists and creating something that works in this particular context. In forms of art emphasizing the distinctiveness of the original less, transformation becomes the focus of the artform instead. In electronic dance music, the songs do sound good by themself but the complete artform only becomes visible when hundreds to thousands of people assemble together in a remote location and set up a festival. In the context of the festival’s (or a techno club’s for that matter) transformation function all the artforms meet and become more than the sum of their parts. And should we then assign the transformation function artistic value itself, we can also see that the festival itself is an artform that keeps repeating and transforming previous festivals to best fit to the current particular situation, and that the repetition of this process is the lifeblood of all the artforms that make it up in return.
The process above is healthy when the transformation function adds value. In order for AI art’s transformation function to add value, similarly there needs to be a connection where it feeds back to those participating their art to the transformers (pun intended).
AI art is a formalisation of the transform function, and pretty much that function only. By itself it doesn’t necessarily add or subtract value, it is merely a function. The system this function plays part in, defines the value transactions.
————————
Afterthougth 2 : The transformation functions of Youtube, TikTok and other autorecommender systems
Alphabet’s (Google) Youtube representatives typically tend to claim they should not be liable for the content they show, because they are simply allowing people to access the content provided by each other. This would indeed be the case, were it not for one important thing: Youtube provides recommendations and pushes them to the end-user without interaction by auto-play. By statistics, users consume 99,7% of youtube content through recommendations or auto-play.
Since Youtube is not simply of storage of videos but an auto-recommender system, in terms of it’s operation as a transformation function it has little difference from what a deejay’s value as an art transformer is. Yes, the deejay is playing other people’s records just like Youtube & TikTok are playing other people’s videos, but nevertheless the deejay is usually held responsible for what the vibe in the venue is like.
I find the above analogy to be useful when discussing what social media platforms do, because similar transformation functions are also at play on the endless scrolling pages of social media apps like Twitter and Facebook. They also display other people’s content, but since they control the transformation of that content into streams with high precision by their algorithms they are assuming the position of the artist creating a transformation of the original artwork in the process, because the output is not the original artworks themselves, but a collage of them. The collage itself contains value, which is in the order and choice, a new story told by sequencing the originals anew, applicable to the particular situation and moment of the viewer.
Our cultural blind-spot for transformation in art may actually explain why we also have difficulty grasping what the role of social media giants really is, and how they are affecting our societies on a daily basis. Possibly our sidelining of the transformation function in art in favour of the original has also left us with an underdeveloped language regarding the important characteristics of algorithms steering the direction of our social interactions. Our public discourse therefore also lacks the signs and proper concepts to handle them and we are left holding the short end of the stick, unable to put our frustration into words.
Gimulnautti 
One thought on “AI art is a remix : The DJ’s of pictures”