AI ART: THE GOOD, THE BAD, AND THE GREY AREA

RISE OF THE ART MACHINES AND WHY YOU SHOULDN’T WORRY (YET)

12/23/2022 at 4:15:43 PM.  This is the time and date that survivors will look back in history and know that Vestige Concept Gallery sold two pieces of AI Art.  The Artpocalypse then ensued.  VCG will hereby be remembered as the Wuhan Wet Market of AI Art.  Does this spell the end of art and creativity as we know it as the mainstream media has sounded an alarm?  No, it’s not (quite) doomsday (yet?)  Worst case scenario, just blame Steve Huth.  (Steve Who-th?)    

Meet Human Steve 

We first started working with Steve Huth in July of 2022.  Soft spoken and fairly unassuming, Steve is a photographer and retired Chief Information Officer (CIO) of Carnegie Mellon University in Pittsburgh, PA.  His prior work with the University also took him to Qatar, where he worked a nearly half-decade residency as the Director of Q-CERT, a collaborative program bringing information and cyber security to that country.  His interests in technology and human collaboration would eventually lead him into the end-user pilot phases of DALL-E.  For those unfamiliar with DALL-E (and DALL-E-2), they are deep learning models (machine learning based on artificial neural networks) developed by OpenAI to generate digital images from natural language descriptions, called “prompts”.  In other words, AI Art.

Prior to Steve, we had no idea about OpenAI, nor had anyone tried to show it at the gallery.  In November 2022, we curiously invited Steve to show his AI Art in our holiday show, and had conversation with him about what all of this means.  Little did we know that less than a week later, OpenAI and ChatGPT became somewhat of a craze in the media, prompting many to say the world was about to end.  So we stuck Steve’s AI Art (very nicely made into quartz- paperweight-esque sculptures) on our display unit to see what people’s reactions would be.  As time rolled on, our swirling questions would become:  “is this total BS?”, or, is this the real-deal?  Let’s find out (and we’ll leave Steve out of the latter half).  

Steve Huth with camera
Huth, rarely seen without a camera. Note: the nefarious world-takeover smile.
Guilty as charged.

The Hall To My Oates?

One of the primary mechanisms (tech-talk aside), of AI “Art” , is that it relies on the inputs from their human masters, or, as Steve describes it: “collaborating”.  Throughout history, collaborations have formed an integral part of the world of art and in music and cinema, collaboration is near-inevitable.  Whether one-on-one collaborations between artists, or studios full of artists, the rubbing together of minds and talents has invariably produced unique works, as well as furthered a greater culture of thought and creativity often leading to movements: a larger shift in both artist and public perspective. Part of the current fear is that AI Art is forming a new movement that will leave traditional artists in the dust and audiences deceived and confused.  For others, AI Art has been here in some degree, and is here to stay (just be careful how you feed the machine).  In a basic summary, Art Fervour says:

“Artistic collaborations are not just based on mutual benefit, but also trust and respect.” 

For better or for worse,  two or more artists cannot be forced to collaborate with one another.  Sure, there may be a mutual benefit (multi-million dollar payout for two artists who hate each other but will reunite to go tour again), but there must be agreement by at least both parties to engage in the process. This is where AI (and their so-called “collaborators”) are disconnected from the outset.  Were it “true” AI (with thought and freedom of choice), we would first ask it (or vice-versa) “would you like to work on a new painting today?”  It is not to say in the one-sided arrangement that some AI “artists” are innocuously well-intentioned, casual users, or a jockey like Steve who puts the limits of AI to the test, but it is that on a fundamental level AI Art is transactional, and there is not a “who”, but a “what” that you are transacting with, and “why”?  The “what” (the technology) has no real say in the matter as much as an ATM, and appears as another tool that serves a means to an end: whether for the human that wants to pat itself on the back (i.e. “I’m an artist now!”), have a little fun (akin to pulling the lever on a slot machine), or to potentially make money.  However, this assertion also leads us to deeper questioning: i.e. are you collaborating with it, or is it collaborating with you?  How much do we know about it?  Will “it” outlearn you, and you no longer serve any purpose?  For now, things appear harmless enough, and we came to learn that there are some known limits to OpenAI, to which Steve must fill in the gaps. 

Assessing AI Art and Transactional Collaboration: The Grey Area

After assessing Steve’s work and by his own admission, DALL-E has its limitations. From the ground level, sites like OpenAI do not permit users to input malicious or harmful prompts, and there are certain safeguards regarding things such as realistic Deepfakes, sexploitation, and creating material for the purposes of defamation or blackmail.  This is not to say that this could change, another reason opponents of AI Art are arguing that we are standing on a dangerous precipice. However, in its current state, digital 2D non-video art is also limitless in these areas (i.e. Photoshopping).  Real-world damage can already occur if manipulated digital art is used in an improper context.  Other limitations as evidenced, are that the AI has a limited (but growing) pool of resources to pull from, as we see in his (now sold) work Christmas Eve: Golden Memories“.  

4″ x 4″ AI/Digital Hybrid Courtesy of Steve Huth
 
Being that our Gallery and Steve are both in/from Pittsburgh, this work is a relatable case study into the limitations of AI.  The initial strangeness of this image lies in its “quasi-Mandela Effect” (e.g. false memories).  If you were born or spent any amount of substantial amount of time in the Steel City, you are without a doubt familiar with the angle of this image.  At first (and at multiple) glace, it is that stereotypical image of the city and your brain immediately tells you that it is Pittsburgh.   However at closer inspection, many of the buildings are misplaced or are combinations of two or more other landmark buildings, or are missing altogether.  
 
Hell with the lid off

 

To give the work his own personal touch, Steve also added the yellow/gold Christmas tree to the image.  This is where we enter further into the grey area between artist/collaborator and what we can define as art and credit for the final product.  Where, or to whom should we give credit?  For some who are extreme-egalitarian: does credit even matter?  In art history, when we don’t know the artist, this becomes known as “unattributed”.  Yet, we live in the modern era where ideas and money do matter to some, if not most people. Can Steve attempt to copyright this image with parts known to be “drawn” from other sources?  How far will Steve go to pursue his creative passions with AI, and, can he profit from it?  

In fact, Steve was so moved that he created an entire series that pays homage to Pittsburgh, with many of the missing/AI-limited pieces being added either through Photoshop, and/or through the use of his photography.  Here, we can conclude that Steve is an artist with a process, and assistance provided by AI, but not a collaboration by the standard definition.  Furthermore, without knowing which parts are which in Steve’s work, we argue that it should be listed as “Steve Huth/Unattributed AI”.  We go further to say that all purely AI Art should be listed as Unattributed.  Case closed?  As far as copyrights are concerned (and returning to collaborations), so much of art and music is pulled unintentionally (or intentionally), from other sources.  The notion of “inspired by” also comes into play, which we see obvious (and admitted) elements of Maxo Vanka, Teenie Harris, et. al., in Steve’s work.  Maintaining copyrights or trademarks can become sticky legal issues, but only so in the sense that you can prove without a doubt another person or group is causing you detrimental harm financially or reputationally. 

“Celebrating Pittsburgh” 24″ x 20″ (Note: the “a” and i” are darkened by the artist for emphasis). Please Inquire for Availability.
 
In opening this article, we used a screenshot image of Arnold Schwarzenegger, from the film Terminator 2.  How fast and easy was it for us to access this image and give zero credit or money to the multitudes of artists, writers, Special FX artists, and the like who worked on that particular scene?  Like a quick Google search or a meme generator, AI Art is extremely convenient when you want something at the push of a button, and it is doubtful that those who become addicted to it’s speed and cheapness will ever let go.  Pandora’s Box cannot be closed, so to speak, and there are multitudes of accomplices and fake usernames, impossible to prove.  Much of the same could be said for the double-edge sword of technology in the Graphic Design industry.  Putting Graphic Design aside, it doesn’t take a lot to imagine how difficult it would have been for us to obtain and publish that Arnold Schwarzenegger photo without the technology.  Without it, we would not be able to get our idea across, while also providing some comic relief.  In the past, a person would have to own a VCR, rent the move, take a picture, wait to get it developed, and so on and so forth… and not many would want to go back to those methods nor can we.  Stealing has become highly ingrained (and accepted) into our arts and communication culture.  Many of the images created by AI happen so casually, so quickly, and so unnoticeably, that it would almost seem completely impossible (and expensive) to prove everything, unless done at an egregious level.  Can an audience even recognize the Terminator provided by AI Art, and would the effect have been the same?  
 
I’m looking for Claude Monet

 

Interlude: Mr. Allen’s Prize Winning Pig  

Somewhere long after Mrs. O’Leary and before Artpocalypse and Steve Huththe future will (maybe) remember the infamous “Jason Allen AI Art Prize Incident,” otherwise known as “the time that guy won an art prize with AI”.  However, this may in the future become the art equivalent to what Fort Sumpter is to the American Civil War.  Shots were fired.  Fortunately for Jason Allen, nobody died.  

For those unfamiliar with this incident, “artist” Jason Allen (of Pueblo West, CO) created a work of art that somehow fooled the bespectacled judges at the Colorado State Fair.  Gosh darnit!  How did that happen?  Furthermore, how did the art community ever let this become such a viral sensation?  While this DUNE meets Bosch work of AI certainly was enough to fool the judges and spectators, we should consider the time and venue in which this was happening.  With all due respect to the Colorado State Fair, this isn’t exactly the world’s standard-bearer for fine arts and culture when also considering the other works of art on display and other categories of competitions that were in the Fair, such as the livestock competition, and the rodeo.  (Don’t get us wrong, we love rodeo).  For others including Allen, they would argue that calling this an “incident” at all is making a bigger deal out of something that to them seems like a normal activity.

But let’s get to the point:  If the artist wasn’t so clearly ripping off Hieronymus Bosch’s “Ascent of the Blessed”, maybe it would have seemed much more apparent to the so-called “judges”, who we now can conclude know nothing about art.  Allen goes so far as to rip off Bosch, that if you look closely enough, you will see the silhouette of the little human person.  Maybe in Allen’s case,  it is a metaphor for artists getting sucked into the endless wormhole of AI Art.   If this is how fine art is judged and subsequently picked up by the mainstream media, than honestly, we (the people) deserve better.  However, this just part of the larger issue: the audiences and media don’t know any better.

Look Ma! First Prize at the State Fair and this $2 ribbon
Hieronymus Bosch “Ascent of the Blessed“ c.1505-1515

 

Fool Me Once…

Let’s face it, nobody likes to be fooled.  Let’s also face that throughout history there have always been fools and those (like the State Fair Judges) who have seemingly been made into fools.  To be more blunt: bad art and artists, and bad ideas have been around since the dawn of time and will never go away.  There’s also no substitute for bad taste, especially once it is commercialized and marketed as the real deal.  What can you do about it?  Nothing, just ignore it.  Only in this Internet driven generation are so many casually forced into caring about what another person is doing.  The more fuel that you throw onto the fire, the larger the cloud of smoke- and that’s when the crowds begin to show up.  Once the building burns down however, we discover that there was nothing really there in the first place (i.e. talent). In a world where people can seemingly call themselves whatever they want to, why should we bother to acknowledge it and to care?  Who is the lone arbitrator on who is an artist or not?  You cannot stop large masses of people from embracing and both using and purchasing AI Art, no more than you can stop a crowd of people from attending a bad concert (that they think is good).  What’s the workaround to this (and in life): style.  Secret: It’s also the workaround to AI Art.  Our answer is to create a style so impeccable, that AI cannot compute or keep up fast enough.  Or, just don’t put your stuff on the Internet.

“OUR HOPE IS THAT DALL-E WILL EMPOWER PEOPLE TO EXPRESS THEMSELVES CREATIVELY”

This statement comes in part from the OpenAI Mission and note the lack of the word “art” or “artist” or “collaboration”.  Ambiguous statements such as these reek of the overly egalitarian virtue signaling that has become a vacuum for mindless Liberalism, while a rallying cry for Conservatism.  Politics aside, traditional artists (and conservationists) who shun AI by definition, fall under the realm of Conservatism.  After all, is Conservatism not merely a way of clinging to the past or past way of life?  Brick and mortar Art Galleries would seemingly fall into the same playing field (and oh, by the way- just think of the children!)  If AI Art is bad taste and classlessness, consider that “to show class”, there must be class.       

The dual fallacies of the Open AI DALL-E2 statement are that a.) “empowering” creates real or actual power among its adherents, and b.) that “creative expression” is synonymous with art, and that everyone’s creativity should be treated equally.   We would argue here that empowering does little to no good (true empowering comes from the Self) by creating a false sense of disillusion under the guise of data gathering, and that the equality/flattening of creative expression only continues to widen the rift between the “haves” and the have nots”, which ironically, seems to oppose the so-called Liberal Agenda.   Meaning, “real” art will continue to rise in value to be traded in less frequency among the rare elite, and the middle class of other real artists (the Artist-Proletariat)  will be forced to slough it out with the “creative expression” folks who are simply poor, lazy, and dumbing down the profession and art values, additionally, cheapening the experience of fine art and the perceptions of art consumers.  As an Art Gallery, our role is to serve as not only an experienced spokesperson, but as a bastion of what good taste “could” (and shouldn’t) look like.  This is also not to say that in rare instances, AI Art shown at an Elite Gallery won’t be sold for exorbitant prices.  It will – it’s inevitable, and the media will give it the hype long enough to create the same “hope” that OpenAI breathes into their mission and followers.  

Sorry kids, we all don’t get a trophy (But you could win a ribbon at the State Fair).  

Continue ReadingAI ART: THE GOOD, THE BAD, AND THE GREY AREA