Andrew is back, and this time, he chats with Steve Coulson about his comic Summer Island, which used AI-generated imagery from Midjourney.
COMIC BOOK YETI: Thank you for making time to discuss your recent work today, Steve. I hope all is well in New York! I came across your online title, Summer Island, by way of a Facebook group called Cartoonist Kayfabe Ringside Seats, where it was shared from your initial post on the Midjourney AI Facebook page, which garnered hundreds of comments and shares on the day after you shared the material on August 18th. What are the most engaging responses you’ve received, and how does this reaction compare to other material you’ve released through social media in the past?
STEVE COULSON: So far it’s received quite a polarized reaction. From the technology community, and specifically from those who are already using AI to generate imagery, it’s received a very positive response. I was one of the earlier users of Midjourney and I think my comic was one of the first ones to make a splash, but I’ve seen many more people creating them since. So it’s definitely gaining in popularity as a method of sequential storytelling as opposed to just single images. And it’s allowing people who did not have the artistic talent to draw their own comic to start putting stories out into the world
On the other hand, it’s received a LOT of criticism from traditional artists, not so much because of the comic itself, but because of what it represents - a technological advancement that threatens to change the economic and artistic landscape. I really do think we’re at a nexus point where AI is able to generate visual images indistinguishable from photographs and able to illustrate in virtually every style (and after every artist) you could imagine. So the conversations with artists have been more about the ethics of the technology as opposed to the output.
CBY: I’ve been using AI image generation to generate tonal reference material for my own art over the last few years, starting first with NightCafe Studios, entering the OpenAI beta, signing up for MidJourney, and frequently using Craiyon. Can you tell us a bit about your investigation of these platforms, and how you decided to use Midjourney? Once you decided on the platform, what has that process looked like in terms of total image creation volume, selection of prompts and tweaking through iterations, and the sort of time/cost involved in generating every image required to tell the story of Summer Island?
SC: I’ve tried a few different platforms now, but Midjourney is my favorite as I find it more “vibey”. Which is a hard thing to quantify, but I definitely think it’s becoming an art culture. Not to knock Dalle, but I have said that Midjourney is a community of people excited about making art and Dalle is a community of people excited about making stock photos. I went for the $30 a month option, which is basically an all-you-can-eat buffet of generations as long as you don’t mind waiting a few minutes for each one. Which is probably just as well because in my first two months I generated over 12,000 images! For the comic itself, I’m guessing it took about 1000 generations to get the 30 pages of art.
CBY: I understand Summer Island as an experiment in storytelling within the constraints of exploring a new artistic tool for visual media. However, on the sensitive topic of resource allocation and supplantation of labour, as someone who has created a piece of work refined and accurate enough to your creative vision to feel comfortable sharing with the world, can you share some of your thought process around what limitations you came up against, and why you may prefer a creative dialogue with a human artist in future comic-related endeavours?
SC: Haha - that last part is a big assumption because I work with art directors and designers every day in my job. Now it’s true I didn’t work with a comic artist or pay an artist to create the comic (because this was done on my own evenings and weekends and the only thing I spent was my subscription to Midjourney) but that was the point - I wanted to see if it was possible, even with the tool in its nascent state to create a somewhat convincing forgery of a comic book that looked like it had been drawn by human hand. SO AI was the whole reason for it being. And look, as I say in the comic “and this is the WORST it will ever be”. We’re on the verge of a massive change in the creative industry, bigger and with wider consequences than the creation of Photoshop or desktop publishing.
CBY: Noting that Summer Island is hosted on the site for Campfire, your multi-media campaign firm, I’m assuming you have plenty of experience building and working with creative teams. How do you anticipate AI image tools and procedural machine learning to support, or supplant, creative roles and functions in project development? When executives can rattle off a handful of buzzwords about the aesthetic they’re seeking and a computer can nearly instantaneously and inexpensively generate something tonally accurate without any protestation or delay in providing revisions and refinements, what do you see this doing to transform creative industries?
SC: I don’t know about “executives” but we’re already using this tool in our workflow - concept art, illustrations, storyboards. Especially in ideation, it’s about visualizing things quickly, so now I can pull all the swipe I need without stock photos or google searches. But that’s a really short term implementation, and kind of masks the big picture.
I hate to say this but AI will affect the production of the visual image the way CGI affected the model making industry. But moreso. If a storyboard artist is worried that directors may no longer need to hire them to create boards, I would say that’s not grasping the scope of the change that is coming soon. The director should also be worried because the natural progression is feeding a script into AI and getting a rendered movie back. That’s a LOT closer than people imagine.
CBY: While clearly Robin Hardy’s 1973 film, The Wicker Man, is the foundational inspiration for Summer Island, there has been the Neil LaBute/Nicolas Cage 2006 remake, Robin Hardy’s 2011 sequel, The Wicker Tree, and a spiritual successor in Ari Aster’s Midsommar, uniquely, you added a crucial dose of kaiju right at the beginning with a reference to an alternate version of the Bikini Atoll test giving rise to monstrous sea gods. What led you to incorporate this element, and how did that impact the imagery you were trying to elicit in later pages?
SC: To be honest, the story is really about Artificial Intelligence and the art community. There’s a small isolated group of crazy devotees, worshiping a monster. And ultimately they feed their new God a professional photographer, a sacrifice they make for their beliefs. And that’s a clumsy metaphor for where we are in the AI vs Art debate right now.
CBY: Digging beyond the overarching references to achieving with Midjourney what you’ve envisioned in your mind’s eye, can you elaborate a bit more on a few panel-specific prompts (such as the cover portrait, the full-page splash on page 6, the image collage on page 14, etc.) and how you were able to get what you wanted to complete your composition?
SC: The prompts for this story are mostly out of date now because even in the time since I launched the comics two weeks ago, Midjourney had a massive update and produces completely different types of imagery. So I doubt I’d even be able to replicate those exact images again. But generally my prompts used descriptive phrases like “black and white line art”, “comic book panel”, “desaturated grey wash” etc. If you are looking for a very specific art style you could also say something like “in the style of Alex Toth and Wally Wood” to get heavier blacks. But even though I did list a lot of references of artists along the way, when I forgot to do that, I still got pretty much the same results with those key phrases.
CBY: You mentioned your layout and editing efforts - obviously, there’s an inherent artistry and skill in composition to structure and pace for the delivery of a story such as this. How much time did you spend “in post,” sifting through raw images and selecting what you wanted to include that best captured your intent?
SC: I laid out the story as I went along using Keynote, which is very user-friendly for comic layout. I knew I wanted to keep to a 6 panel grid structure throughout so I used combinations of that aspect ratio in every image - 2:3, 1:2 etc. And as i say, I generated a lot of images. Some came almost straight away, but then the last 10%, ones that were key images in moving the story along, took 20 or 30 rerolls. I lightly used photoshop on maybe 10% of the images. The cover is a comp, and the page where the photographer gets knocked out is comped. And the photomosaic was created with a mosaic generator. But mostly what you see is raw generation from Midjourney. Fonts by Blambot.
CBY: Does your experience with Midjourney parallel your experience with any other computational platforms or novel technology? Having utilized this tool extensively, what sort of ontological or ethical questions, if any, does it raise for you? What kind of responsibility do human beings have for guiding the cultivation of a benevolent AI, and what impressions does it leave you with around the idea of singularity and the advent of a synthetic intelligence accelerating beyond the capacity of human thought?
SC: I’m just a guy playing with a tool, and a tool that’s so advanced, it feels like magic. But I’m not the philosopher you’re looking for. I would say there’s no such thing as benevolent AI, there’s just AI, it’s what good or bad people do with it is the issue. There will be a transition as there is with any technology. iPhone put Kodak and Polaroid out of business and I never would have seen that coming, until it happened and you go “of course it did”. But I wouldn’t expect Skynet for at least the next few months
CBY: While Summer Island’s inspiration is quite clear, what other media would you cite as an influence on your creative process? Which comics, films, and other media are most fundamental to how your aesthetic worldview has developed?
SC: All comics. As I say in the back of Summer Island, I’m a huge fan. I started reading comics regularly in 1969 - I still have those issues - and have probably read 10-20 titles every month since then without a break. I have sixty boxes in my attic and I stopped buying physical years ago. So despite some comics artists being very hostile to me on Twitter about this, it’s still a love letter to sequential storytelling.
CBY: In a final moment of appreciation for our readers, can you share with our readers which artists/creators’ work you’re most enjoying these days? What out there would you consider required reading/viewing at this moment in time?
SC: I’ll buy anything and everything Tom King or Ed Brubaker writes. And I wish Chris Ware would start drawing again.
CBY: Thanks for making time to field these questions, Steve! Please feel free to share any social media or other pertinent links and resources you’d like to include for the benefit of our readers.
SC: Please go read Summer Island - it's a free download and available in YACreader friendly CBZ format as well as PDF - https://campfirenyc.com/summer-island/