Attack of the Artistic AI Bots!

Earlier this month it was reported that an AI bot by the name of Midjourney took first place in art competition held in Colorado, quickly creating a storm of comments spanning the spectrum from welcoming our new benevolent overlord artists to the downfall of society (or at least art) as we know it.

No matter your take, the world of Art just heard clearly the clarion call of a revolution as great, and likely as disruptive, as the advent of photography. It was only two centuries ago that the photography was introduced to the world, and it would not be till the 1940s before it was seen as an artform. A similar journey awaited movies, and now video games.

The groundwork for this revolution has been quietly happening underfoot in some of the least likely, read: geekiest, of places such as at Google’s AI labs amongst others. While OpenAI’s DALL-E has been out and available for awhile, we now have any number of publicly available AI bots that have been created with toolsets that make them significantly more accessible to the average person.

More so, they are being commercialized as we speak. Midjourney, which made public its beta on July 12, 2022 – a scant two months ago – is in talks to raise capital at a $1billion valuation. While Midjourney has made the headlines, competitors are not very far behind with Stable Diffusion from CompVis along with a handful or more of AI generating image tools gaining traction.

Rise of the machines

But I’m not really focused on talking about the latest AI bot tools as this is changing, if not by the hour, then by the day presently as new training models, updated and improve toolchains, et cetera are developed and released. Regardless of whether you are a “glass half full” or “half empty” person, the last few weeks have been quite a white-knuckle ride.

In a lot of ways, these recent advances mirror the same ones I saw back when Adobe Photoshop and other digital photo tools become available for the Mac and then PC. Albeit the pace of innovation in 2022 is orders of magnitude faster than it was in the mid and late 1990s. Nevertheless, it’s the same conversation. It was said then that “real artists don’t use Photoshop”. Nowadays you cannot find amongst popular artists who do not include some amount of digitization into their toolchain. And I suspect will live through a period of “real artists don’t use Midjourney” nonsense.

Goofy faces in style of Pixar

If music videos killed the radio star, then there are people who say digital art killed the fine artist. But the reality is that none of this true. Yes, there are people who will always prefer the value of a physical object, as evidenced by art galleries the world over. My point is not that more traditional artforms are being replaced by digital, but that digital is maturing to be on par and peer to these other forms of art. The sale of Beeple’s work for $69 million at Christie’s is proof that there any people in the world putting “their money where their mouth is” when it comes to this thesis.

For myself, as an artist, worse than the doomsayers are those who have the opinion – both within and without the art world – that digital art, and now by extension art created with AI, are not real artists or real art; that somehow the use of more advanced tools diminishes the essence of the art produced. But if that is the case, then the whole of history since our ancestors put art on cave walls is so reduced.

The simple reality is artists throughout the ages have leveraged whatever tools were available to them to improve their craft. Just look no further than the camera obscura and camera lucida. There is mounting evidence shows that many masters of the time extensively used these tools. I would further argue that the use of these tools improved their work, minimally at least for those who did representational, perspective pieces.

Camera Lucida used by the great masters of the 18th and 19th century

I wrote on this before about art as artifact versus art as process, but too often we do not separate art as a consumptive experience from the creative experience. Which is to say, when we talk about “what is art?” there is rarely an explicitly shared common understanding of what qualifies as art. Worse, we often conflate the qualities of art, or aesthetics, with the art itself. We are all our own Justice Potter Stewarts when it comes to the definition of art: “I know it when I see it“. Which is really just an admission that art is subjective. Has been. Is now. And forever will be.

Just yesterday, I was chatting with Mark Ferrari, my good friend and artist/author extraordinaire, about this post’s topic. And while I may butcher what he said more elegantly to me then, I subscribe to his notion that:

“the essence of art is to take the internal and make it external so that others may experience it as the artist experiences it in their mind.”

Mark Ferrari (paraphrased)

That’s it. Everything is mere window-dressing, or more specifically technical aspects of the craft to make art.

Whether you use traditional processes and mediums such as oils, gouaches, watercolors or use, as I do, digital matters very little to whether something is art or not. These tools do not add to or detract from the value of the thing created. What really matters, and who is ultimately an arbitrator of the goodness of the art, is primarily the artist and secondarily the viewers of the art. Primarily the artist as only they know if what they created is true to the imagined internal concept they had. Secondarily the viewers as they can only affirm or deny whether the art they experience conforms to the expressed ideas of the artist. Note, you can invert the priority of primary and secondary, but then this is difference by degrees but not of type.

The faces of our new overlords?

Based on this definition of art – which I believe is the only honest definition of art one can make – then the advent of AI bots to generate (or assist) with the creation of art in no way influences whether something is art or not. Or whether someone is an artist or not. To be clear: art itself is devoid of relationship to the means of its creation. The only question then is if the tools enhance or impede the artists ability to meet their own intent.

I do not foresee a world where art loses its meaning with the advent of these tools; quite the contrary, we will see art continue to evolve as these tools, and yet to be imaged one, evolve. Just as it did with the advent of oils. And camera obscura and lucida. And photography. And Photoshop. And now AI-generated art.

The question is not even when this revolution will occur, it’s already happening. We see professional artists learning how to incorporate Midjourney and other AI generative tools into their workflows from game asset creation to creating a graphic novel in a day to proof of concept web design.

The bigger question is to where will these tool take us? For myself, I think these tools will continue to help bridge that gap between what we as artists imagine and what we dare to create. And that is an exciting thing indeed.

I, for one, welcome our new benevolent artlords.

Note: all images in this post are all generated by the author using Midjourney.

Erel

I’m back at playing D&D with a new group of folks, and as such it’s time to create a new portrait for my character. While I considered playing either Stone or Unnis – given this is a more role-play heavy group (versus the min-max tactical approach in my other groups) – I also thought it would be fun to flex into something entirely new(ish). I say new’ish as my sense of role-playing is to play out aspects of myself, so while Erel is new as a character, his traits borrow heavily from aspects of myself. Why? Maybe because I’m autistic, but I find it just easier when I’m not pretending to be someone I am not – both in real-life and in-game, as it were. If you’re interested in learning more about Erel, then I’ve included his backstory at the bottom of this post.

I actually did two pieces for Erel, both a close-up portrait and full-frame view. As you might have noticed, I copied the portrait into the full-figure. I originally started out wanting to draw Erel’s face a second time, but the re-use (aka copy & paste) boils down only partly to expediency, but more to the fact that I’m not convinced I’m good enough (yet) to reproduce an image of Erel a second time and that it would feel convincingly like it was a picture of the same person. So yeah, I cheated.

With Unnis, I play a character who is in many ways closer to who I am today as I near 50. With Stone, I pull from my tendencies toward toddler’ness with his childlike simplicity thus Stone is 3 years old. With Erel, I wanted to return to when I was in that idealistic stage of my life when things were very binary and clear cut, but were ideas lacked nuance that only time and hard-won life experiences can provide. In this way, I arrived at a younger person in what I envision is their early to mid-twenties.

I debated adding wings to Erel. As you can see in an early stage of values Erel has no wings. But as I progressed, I kept struggling to find ways to convey the fact that he is Aasimar. In the portrait, I just added swirling colors and tattoo to minimally convey something ethereal about him. In full figure you’d think that I’d just add wings and call it a day. And while most Aasimar do have wings, how and when they present themselves in context to D&D matters. In particular, Erel is still very much a green level 1 paladin. The use wings, in the strictest sense of the game, does not occur till level 3. And if he is level 3 (he is not) there is debate as to how Aasimarian wings manifest themselves. Actual wings. Skeletal wings. Wings of radiant light. Putting aside all that nuance, or pedantism, I felt wings of some sort were warranted. I happened to be watching Ross Draws while trying to figure out an anatomically correct version of Erel’s wings when Ross said something to the effect, when asked why he had done so and so to a particular picture, “because I liked it even if its not realistic.” So guess what? I listened to wise Ross and just drew what I liked regardless of anatomical correctness. Go ahead and arrest me! In the end, they are just a visual device to convey to the viewer his race. And frankly, I like them so go and bugger off if you wanted to see something from the Auduban.

Like every picture I create, I am both pleased and disappointed in the results. In some ways, I’m happy with the portrait of Erel. It was a relatively quick study, and I think generally has a consistency in execution that makes it feel complete both visually and narratively. On the other hand, I think Erel in full-figure still has places where my choices of where I put hard and soft-edges bother me. For example, I think his left hand holding the glaive needs to be softened some more. And I think some of my values on this left leg’s armor look too flat. That said, I’m happy with how his wings are desaturated and generally only have soft edges, helping frame him as a subject without distracting the viewer’s eye.

That all said, Erel in full-figure is, to a trained eye, a great example of where I struggle as an artist: namely; moving from a place of lines and capturing every detail as if I was a camera (aka photo-realistic artwork – which the technician in me admires, and the artist in me cringes at) to a more painterly style that focuses on interesting shape language to convey the subject, with a love of soft and hard edges to “sculpt the form.” As noted, I think the wings show I can execute on this, but the question next is: can I do it for an entire image? Trust you I, I’m going to continue to push myself as an artist to find my limits. That said, the urge to approach art as if I’m holding a pencil and not a brush is very, very strong in me.

Erel is a bit of firsts for me, as an artist. He is the first character where I’ve done two pictures of which is exciting for me, even if I did take some shortcuts to create continuity between both pieces. So go me for trying my hand at building a portfolio around a single character, which is an important skill if you want to work professionally in games or movies. A part of me secretly holds out at getting a job in “the industry” where such production skills are paramount. And for any of those of you who are already in “the industry”, I’ve been around long enough to know that I’m better off as a hobbyist where I’m my own “art director” than working for “the man” in “the industry.” Enough of “those quotes.”

Another first for me is Erel is also done entirely using Adobe Photoshop between my Wacom and iPad Pro. To put this into context, till very recently I relied heavily on Procreate on my iPad for quick studies, and then I’d move to my desktop with Wacom Cintiq for more control and greater detail as I entered the later stages of development. While this is a fine setup, I’ve struggled with Corel Painter’s flakey UX even when I know it’s brush engine is superior to Adobe. On top of this, there is friction sharing files via import and export via Apple iCloud for syncing between devices. It works, but at times I find myself struggling with the whole workflow. And worse, I like to have options on when and where I can work on my art. This workflow means I lose the freedom of the iPad once I lock myself into Corel Painter. A first world problem for sure, but I really like the ability to be a nomadic artist.

So why Photoshop? Especially when one of my favorite artists (Todd Lockwood) is a staunch defender of the superiority of Corel Painter to Photoshop, which he relegates to hacks. And I don’t want to be a hack! But outside of losing the love and respect of Todd (assuming he knew me, which he doesn’t even though he and I’ve chatted – swoon in my head we are totally BFFs. Hey! It’s only awkward if Todd thinks it’s awkward ) is the simple fact that Photoshop is “the 800-lb gorilla of digital art”, and as such I pin all my hopes that driver support is just better than with Corel Painter and Wacom. Sorry, Todd!

More than just trying out Photoshop, I opted to save my PSD files in Adobe Cloud. At first I resisted this even more than just trying out Photoshop since I did not want yet another cross-device file-sharing app. I already have all my artwork on iCloud and I’m very happy with Apple’s iCloud, please and thank you. However, I’m so glad I did because when I found myself with a twisted knee late last week, there was a few days where I couldn’t sit at my desk and work on Erel. So on a lark I decided to download Adobe’s Photoshop for iPad, having recalled reading it was a pretty good competitor to Procreate. Given I’m a died-in-the-whole Procreate fanboy/fanman/fanperson, I did not entirely believe the hype and was disinclined to then to even try it out. Like, how can you be better than GOAT Procreate?

Imagine my shock and surprise, pleasantly so, when I launched the app on my iPad to discover I could just open Erel and off I went into full artist mode, no compromises. Not only could I do complex changes, not just simple touch-ups, but it was an amazing drawing experience on a tablet that is on par, and in some ways, better than Procreate! Sacrilege! I know, right? But wait, like a great ginsu knife, it does so much more. Imagine my greater shock just a few days later when I went back to my desktop setup and just picked up where I left off. All with nary a hiccup. Color me impressed.

Recall my wanton and unmet needs as a nomadic artist with my old setup? Adobe Photoshop coupled to Adobe Cloud means I truly can work anywhere I want, whenever I want, using whatever device suits me best. I absolutely love the possibilities!

Backstory

Born Erel, First Son of the First House of Temple at Earsa, it was written that he was destined for great things.  Written not just in the stars, but in the very essence and blood of his race of Aasimar, servants to the gods.  But what the gods wrote and what Erel read of it were two different things.

Since childhood he trained arduously as first acolyte then attendant to hem-netjar of the temple at Earsa, overseen by his family for the past seven generations.  While devoted to his studies, his tendency to question scripture led the hemet-netjar-tepi (high priestess), his mother, to believe he would best serve the temple by joining the ranks of the protectors to the temple.  It was hoped he would someday rise to the rank of hem-asim-tepi (high proctector).  In this capacity, he would stand side by side his younger sister, who it was believed would eventually replace their mother as hemet-netjar-tepi.  In this way the gods would be served, siblings securing their family position as the First House of Earsa for an eighth generation.

Following a decade of intense martial training as a protector and joining the temple garrison, Erel showed no signs of stopping his challenge to scripture.  While exemptions were made given his relation to the temple’s hemet-netjar-tepi, Erel struggled to stay silent as decades of dissatisfaction mounted around the scripture’s deepest truths: that all of existence is but to serve the gods, and more so that all of existence — all that was, is, and will be —had already been foretold by the first god, known only as the “the unseen god”.

As Erel’s questioning grew argumentative, less as questions and more as posits, many in the temple feared that his words would come to corrupt others, if not outright bring the wrath of gods upon the temple to destroy it.  He argued heretically that the Aasimar were not servants of the gods, but in fact their slaves.  He argued that the suffering on the lower planes was not preordained, but instead a direct consequence of the indifference of gods.  For Erel, the gods’ place in the greater cosmos and across all planes were not to be served, but instead to serve.

He only relatively recently left his onclave, albeit truth be told, it was either leave voluntarily or quickly find himself facing a tribunal for beliefs and acts unbecoming an Aasamir.  He would be stripped of all rank and privileges, and forced to live a life of an ascetic until he atoned for his sins. His accuser was none other than his very own aunt and hemet-asim-esun (protector, second order) of his protector regiment, where they had witnessed Erel burning sacred scripture while arguing to his trainees that what they had been taught since birth was wrong.

Erel, now marked as a pariah amongst Aasimar, is outcast amongst his people.  He goes only by his given name, having otherwise left behind everything that he has ever known.

He has been wandering darker roads in the Underdark under Waterdeep, believing that his true purpose must exist upon a path only he can divine.  While he has never broken his vows to protect, for it is one of his core values, he has no reticence to harm those who prey upon those weaker than themselves.  In the past week, he has found himself in Waterdeep proper, continuing to seek answers to questions for which he knows to ask but can find no answer. 

Stone

I recently started playing D&D 5e on Camp D&D Online after backing it on Kickstarter. I tend to play magic users pretty much any chance I get, so for a change of pace I thought I’d play a melee character. For point of reference, some 35 years ago was the last melee character I played was a dwarf fighter with big dreams of becoming a paladin. It should be noted that this was in the days of 2e where such things were not, per canon, allowed. And as you might guess it with a bunch of teenagers, the DM refused so I just roll-played as if my dwarf would some day catch the notice of a human paladin who would induct me into the eternal order of protectorates. It never happened. Dwarf prejudice was a real thing, kids.

Before you worry about me, I also created a dragon-born sorcerer by the name of Unnis Kilyax, for whom I will illustrate later this week. But for whatever reason, I started with my first fighter in decades creating a warforged warrior who woke from a scrap pile of their brethren without memory of its past. Stone exudes a child-like innocence that sits uncomfortably with the fact that they are an elite, two weapon wielding mercenary which its named Sword and Axe. Warforged, if you don’t know, are decidedly simple and direct in all things including the naming things.

When I started sketching Stone, I thought I might go with a more straight-forward rendering using line-art and cell-shading. However, I really did not like the line quality, and I started to tweak I ended reworking the entire piece to be a more painterly rendering. Once I got done and let it sit for a few hours, opting to change some subtle shading around the mouth along with details on the face that I think help keep the eyes on the face.

Given that Stone is a walking automaton with a body that is effectively a full-body suit of armor, it did not make sense to have them wear a helmet. Interestingly enough, I hate the hood as an element from the perspective of character design, but I never figured out a better approach. I struggled how to convey that Stone was a fighter in a portrait, thus why I added a sword-like symbol over the forehead as a compromise. Admittedly, there is an error of mystery with the rather organic elements of the hood with what is otherwise an entirely metallic and mechanical humanoid.

Level Up! 2016 to 2018

I’ve wanted to go back to some of my older pieces and re-do, largely born out of both a dissatisfaction with my treatment of the subject and a desire to see if my skills are sufficiently “leveled up.” Ironically, I keep psyching myself out worried that maybe I have not grown enough to warrant the effort. Well, I stumbled on a post from May 2016 and realized I had inadvertently already done just this.

Continue reading “Level Up! 2016 to 2018”

Dragon Doula

This piece is meant to be a part of a series of two portraits. One of myself as Dragon Warden, and the second as a portrait of my most beautiful and excellent partner. Both of these pictures were taken from a single reference photo of the two of use at NorwesCon 38 where George R. R. Martin attended as the guest speaker.

Continue reading “Dragon Doula”