Mike Wheeler announced on facebook he was able to use AI to emulate the work of popular Indian comic book cover artist known for creating hyper-realistic NSFW covers, Shikarii. This news comes as a bit of a shock, being that Shikarii’s artistic process has remained elusive and his behavior surrounding the process controversial, often blocking inquiring consumers who ask about the process. To find out if the hyper-realistic process Mike has uncovered is really as easy as 1, 2, 3, read the full article.
While this article focuses on Shikarii’s work as a case study, the educational information concerning AI can be extended towards all artists and for all styles. The AI tools we’ve investigated are most developed towards either hyper-realism, anime, or content of an adult-nature. Today I’ll be exploring only hyper-realism for comic covers, although there are MANY more AI tools, even some that will let you create your own Anime with ease. Those are great for topics for another day.
Shikarii: A Case Study in Transparency
Shikarii, a former video-game designer who is advertised by Comics Elite as “drawing faces from scratch” and drawing digitally, has exploded all over the indie comic book scene over the past 2-3 years, commanding a high price for his covers. He is most notable for appearing as Comics Elite’s star child cover artist and promoted exclusivity with them for a time. His credits also branch into most every popular indie title including Kamen America and more recently, Lady Death. Die-hard fans claim Shikarii’s work is authentic, while many artists and consumers believe that this is an obvious denial of reality. The use of 3D models, AI, and photo-bashing in Shikarii’s work has been heavily called to question, yet Shikarii refuses to show his full process. He redirects further questions to his publishers, which he says will vouch for him.
While work in progress (WIP) screenshots have been shared, many artists agree these process shots from Shikarii allude the majority of the drawing process, causing them to believe these are nothing but deception. Fueling the contention, Shikari took a firm stance last year when confronted, swearing he would never show his process in video form. Never.
Here are some WIP’s. It’s very important to note the use of black and white lines over already rendered black and white artwork here. While these are meant to persuade us this is not AI, one could equally argue it produces the opposite effect being that it looks very much like the use of Depth Maps. Depth Maps are part of the AI process in certain softwares. We will explore more about Depth Maps further down in this article.
Support from Publishers
Shawn Hudachko from Comics Elite backed up Shikarii during this confrontation last year, sticking firmly to his guns that Shikarii has “proven himself” but failed to provide a valid reason why the process would stay hidden beyond the claim that we should trust Hudachko’s word. Well, since then, former employees of Comics Elite have spilled the beans on Hudachko’s use of AI for writing scripts, despite both Hudachko and Shikarii claiming to be “leaders in the fight against AI.” To make matters worse, issues of non-payment and lack of due credit have arisen including the time Hudachko somehow lost all the money he raised for Artist Elite (an off-shoot of Comics Elite). We can suppose Shawn’s word may just be the opposite of reliable, given that information.
Even SWORN NATION dealt with backlash when announcing Shikarii’s Lady Death cover earlier this year. After many members questioned AI and expressed disdain over the cover, a group admin deleted the SWORN thread and Coffin Comics has remained silent on the issue. Both Comics Elite and Coffin Comics benefit monetarily from Shikarii’s work.
Comics Elite’s ebay listing for Shikarii’s Lady Death cover: https://www.ebay.com/itm/166963342851
3d Models and Celebrity Likeness Issues
To contest the claim of Shikarii’s faces being drawn from scratch, many have traced the same celebrity faces Shikarii uses to easily accessible 3D downloadable models (Jenna Ortega, Gal Gadot, Stormy Daniels, Taylor Swift, Margot Robbie, Scarlet Johansen, Emilia Clarke, Gianna Michaels, and more) even down to the hand that rests on Wednesday Addams shoulders. Here lies another Shikarii controversial cover in which he pictured Wednesday Addams in full frontal nudity, despite the character being an underage girl (he apologized for the cover after the fact). Shikarii’s style also caught the attention of Margot Robbie last year who shut down the Comics Elite Barbie “Cosplay” cover with a cease and desist, costing Comics Elite a pretty penny. Since then, Hudachko and Shikarii have nested themselves into Tidal Wave Publishing where they can use celebrity likeness in the realm of biography without finding themselves in trouble like they did before.
It’s just too good to be true in the eyes of many that a video game designer with these skills would spend dozens of hours painting hyper realism when he could simply download models to photo-bash or input into AI to photo-bash. Yet no one seems to have been able to nail down and be transparent about a replicable process… Until now.
Stable Diffusion
While it’s true, Shikarii’s digital art predates AI, his work seems to have improved since the release of Stable Diffusion in 2022, around the same time he was picked up by publishers. His use of the blur and other effects combined with the pore brush appears to have also increased, suggesting this may be a tactic to hide AI generated mistakes. It’s been possible to use 3d models for comic covers for over a decade now and possible to photo-bash since the 90’s, so his portfolio leading up until the point where he found comics doesn’t negate the suspicion of his current artistic process. You can view Shikarii’s alternative (and highly NSFW) portfolio of mostly transgender erotica (specifically chicks with dicks) from a different Twitter handle (@phutarii). This is important because it sheds a little more light as to why there is such a polarized opinion of his artwork in comics. A lot of his original fan base comes from a world of appreciating the end result of photo-bashed body parts more than the process; whereas, traditionalist comic fans have prioritized the strenuous process of comic book art, sometimes even more than the end result.
Our Findings
As suspected, AI gets very complicated and technical. Both Mike and I having come from a background in web development immediately recognized the crossover of mistake-prone artists and foreign video game designers or web developers from countries where outsourced development work was prominent. Developer jobs are most prominently outsourced to India, Russia, and China, among other countries, so it’s no coincidence you will find many extremely tech-savy individuals there.
For the average Joe, you have your basic Bing Image Creator tools and a little more advanced is the very stunning Generative Fill feature in Photoshop, but I knew it went far beyond that. AI extends into a hacker-like realm where you have to install multiple softwares on your computer and run them on different operating systems. You need to be familiar with coding to understand how to get there. Oh, and you need a massive amount of processing power, a high-end video card, and it can only be done on certain computers. The Average Joe would not be able to get to this level.
If you lived outside of the US and lack the ability to have a high-end graphics card, there’s a work-around for that. You can set up a Huggingface account which enables you to utilize Google’s graphic card processing power via an API. You can then save the results through Google Drive.
The AI Process
If you are tech-savy enough to follow this so far, Mike has done the comics world a favor and laid out the entire process in a Facebook post so you can look it up yourself. To sum it all up – after you install the AI programs and get into the meat and bones, you make a stick figure. It looks very much like rigging a 3D model and technically, it is. You then input a photo of choice (an “adult” photo probably works best). This can also be an AI generated photo. The AI will then generate a black and white rendering of the image called a “Depth Map” which it uses as reference in conjunction with the stick figure, along with any other prompts you tell it to integrate. (More on Depth Maps below).
These tools combined enable you to manipulate what gets generated out of the AI with enough control and detail that you can change the lighting, pose, details, and features. Models exist for you to download which are pre-programmed in certain poses (usually NSFW). Yes, you can even give it Taylor Swift’s face. Once you generate something you like, you can download a high-enough res version and turn it into a nice looking comic book cover with mild to moderate photo manipulation. It took me only about 50 minutes of screwing around in Procreate to turn this AI generated piece (left) into what looks just like a Shikarii comic cover (right) and additionally, we were able to fool the AI Detectors with minimal effort.
Depth Maps
You can even export these AI results into a 3D modelling program as their own 3D models to further manipulate (thanks to Depth Maps), and then go back into the AI again for further tweaking. Or, just manipulate the Depth Map itself and re-input it into the AI to generate a more refined result to your liking. Black chicken scratch lines would easily tell the AI to put more shadows there, as black means farther away and white means closer to the camera. At this point, the process is pretty much just a back and forth between AI and photo manipulation programs until you get the desired result. The possibilities are endless here.
If you’re familiar with laser engraving for something like a metal challenge coin, you’ve seen what a Depth Map looks like (see below).
When you see artists painting black and white “art” and then magically, in the next Instagram post it’s now in full color, it could be that they are actually just painting on top of the Depth Map, prepping it to go back through the AI. It also might make more sense of Shikarii’s WIP’s and why there’s no in-between – just chicken scratch to rendered result.
The Age of Transparency
The point at which anyone could tell whether AI was used in a piece with certainty came and went a long time ago. AI is rampant and has flooded the market already. It’s even been in dispute whether DC and Marvel artists use it. The best you can do is educate yourself on AI and pay close attention to the artists’ transparency and honesty about their work. Work-in-process videos and photos are great but also sometimes made to deceive. I imagine in the near future, even WIP pictures or videos could be generated completely by AI and you could be “art cat-fished” into obscurity. I’d say the only way to wholly judge an artists’ capabilities would be to hand them a blank piece of paper (one they haven’t projected traced lines onto already), and watch them draw live. No breaks, no references, just pencil and paper, start to finish. Many comic artists such as David Finch do this already, daily. That’s his job as a comic book artist.
Shikarii is welcome to show his process or not, nobody is forcing him either way, but in the Age of Transparency, be prepared to face negative consequences if you choose to be secretive about your process.
AI Consequences
The good news here (Artgerm and I both agree on this) is that in the end, AI only increases the value of human-made art and that desire will continue forever. Artgerm calls himself Artgerm because his art is like a germ – spread everywhere. As one of the most copied artists out there, he has succeeded in sharing the joy of his work all over the globe. Still, if you want an actual Artgerm piece, you’ll need to go to Artgerm for it. AI can be thought of much the same way. If you want human-made artwork, it is still out there for you to find and appreciate. The oversaturated AI market seems to only make human-made art more appreciated.
Can I Support AI?
Why not? You can support whatever you want. Although AI is extremely controversial I have a pretty lax attitude towards it in general and my opinion is ever evolving as I’m continuing to learn more about it every day and the people who use it. Having confidence in my own artistic abilities, I don’t personally see it as a threat and I work hard to improve my own artwork as much as possible in order to compete with AI artwork on the market. I enjoy the challenge. This article is only here to inform of the social issues within the realm of indie comics and not persuade you one way or another.
I’ve hoped you enjoyed what you read and will share it with friends and family.