a response to Jonas Čeika – CCK Philosophy – “What would good AI art look like?”

I did not seek this video out, it was shared in a discord wherein I interact with various friends and acquaintances.

I have written a response. perhaps due to my tone and word choice it may read rather, as a retort.

I’ve written several mini essays previously where I share my thoughts on this subject, similarly off the cuff and haphazardly formatted. To encourage elaboration and examination amongst peers, and for the sake of my own sanity, and ‘time’

[a soliloquy, or uncharitably a tangent]
imo it’s a pretty shallow take on the capabilities/benefit of A.I. Image Generation. The comparison to photography is apt, and not without merit with regards to any kind of Media we create, and when it comes to comparisons between methods and mediums. however. Good photography takes time, requires a certain understanding of specific framing+lighting+etc concepts. But these are things one can learn, and the barrier to entry is lower than ever (I gotta get out my Vita and practice some Vita Photography some time /hj)
Personally, prior to the advent of ai image generation I have levied many haughtily self righteous complaints one may have about ai image generation in regards to “Virtual Photography”
given how the ease of access ‘in engine’ tools many games have, provided many the ability to produce images akin to what used to be regarded as a moderately skill based hobby. My implicit acknowledgment of the value and personal appreciation for the process/tools turned to disdain as spaces established for artworks came to resemble one’s PS3 screenshot gallery. My arguments towards them are much of the same. They did not create the assets contained within the image, any credit given should be given exclusively to the original artists. Even with regards to the process of choosing framing, lens measurement, post production filter, etc which nowadays in many cases mimic or are even synonymous with traditional photography, do not often produce the same results/quality that a traditional photo could. an intricately high level of personal investment is required to produce images like any example one may refer to as exceptional photography/graphic design.
or even just any base line example of artistically inclined photography.
to that end, I could imagine A.I. image generation might garner the same eventual plateau of engagement/proliferation as ‘Virtual Photography’ a technical curiosity with a low barrier to entry and any manner of wide ranging potential one could place on it, for every 1000th static dull uninformed image easily reproducible by anyone else with 10 minutes of free time, there could possibly be actually decent and engaging images that owe their detail heavily towards generative computer scrapbooking. Not unlike an intricately “photoshopped” collage, with the end result so strikingly distinct from it’s source that it reserves the right to labels such as ‘transformative’ or ‘kitbash painting’ but the sticking point for me – is that much like with virtual photography- A.i. image proliferation creates a congestion that forces a significant lack of visibility of more traditionally produced/styled art works. as we have observed just on twitter in the last year or so, for every one drawing/painting produced by a traditional artist, an a.i. prompter will publish 20 similarly framed, increasingly mediocre (by way of overexposure and repetition), highly flawed (explicitly for a lack of trying) images, and the publishers of such medias smugly grant themselves an air of self satisfaction not even many of the most prolific illustrators you or I follow feel themselves with regards to their own art. Though not stringent morally, these publishers evoke dirty looks and barks of disdain very similarly to ones I’d given to virtual photographers, perhaps moreso even – due to their self imposed positioning in place of and in venomous spite of traditional artists, as they make claim to the images as “their creations” as though every line/shade/curve/and colour implied in the work were chosen deliberately by them.
they are not ‘virtual photobashers’, that isn’t sexy enough, that can’t be put on a book and sold at a premium one thousand times it’s worth to the unassuming prospective buyer, and it can’t be used to sell generative tools to potential customers. They prefer the deceptively plainclothes moniker of ‘a.i. art’ if you can even get them to admit that it is an image that was produced generatively.

after all, what is “Art” to the average person?
Art is a Painting, well the prompter says “here is a painting, just as good as any other. No, even Better.”
Art is a low poly dutch angle screengrab of Sonic in the City Highway level from a position not possible within the limitation of the playable environment.
well the prompter says, here is that, and they do so with much aplomb. but the polygonal edges are blurry, and the colours are wrong.
Art is an illustration caricaturing a popularized or iconic subject for the purpose of affirming appreciation, or to satirize.
to which the prompter proudly claims to be able to produce infinite variations to the form, severely lacking however, the finite (as it pertains to the goals of the individual and resulting executions of those ideas) yet humanly paradoxically infinite potential intents of a deliberately designed work.
and so they seek and are rewarded, socially, and financially for a media(s) that they owe explicitly the effort of another, many others, with little to no clear path to or even recognition of the original artist’s efforts.
If I take a picture of a scene in Kingdom Hearts, everyone knows, even if it looks good “that looks good, and it is Kingdom Hearts. I recognize that structure, I recognize that this is the npc model for Scrooge McDuck”, if I take a picture of the Seine River, I have informed all to whom I share it with, “I shot this at the Seine River”, a photo of the Sky, or a photo of my Vita, is recognized as such and appraised on the merits of how well I captured the subject, and what details I chose to highlight, and what tools and methods I utilized to do so, distinguishing it from a photo anyone else could have taken of the exact same thing – etc
this is what brings about the aspect of the conversation that has become shorthand due to the exhaustion the discourse has produced in such a short span of time, that aspect being- if we are to take generative a.i. image producers on the merits they are publishing under, that merit and categorization being
“this is a drawing”, “this is a painting”, “this is your mom, but with enormously impossible tits”,
their work is produced and proliferated by way of theft.
providers of the software that receive income from offering tokenized services to produce these images do so as a result of theft.
consciously and actively the designers of the most effective and popular of these services built (or rather “trained” as they like to claim) their a.i. on the work of Artists, specifically to recreate their images, without any means or method of crediting their originals. tantamount to theft.
we cannot accept or allow a.i. generated medias a place, even within the dingiest of our galleries. Because of it’s overt application of deceptive sourcing, and deceptive publishing. Not even on the backs of the idealistic fantasy of an ethical “a.i. artist” (of which I have spent many hours debilitating and brainstorming the possibilities and potential of) can we resign, as it still leaves room for that ever looming theft and deception. Even at least, a Painting Forager has to know how to paint first. We can forgive, this transgression, in the hindsight.


a.i., I feel – we should not.