I hate writing, I love having written.
Dorothy Parker
The emergence of online tools enabling users to create images from text prompts has sparked debates about generative AI, resulting in polarized opinions and ethical dilemmas within the cultural industries. It is within this crucible of contention that comics, a medium that has depended for the greatest part of its history on well-documented craft tradition
s and materiality signatures, emerges as an emblematic case study. The increasing access to these tools urges many industry insiders to express a deep-seated unease regarding AI comics and the impending impact of these nascent technologies upon the established craft traditions, the thriving ecosystem of comics industries and the delicate equilibrium of fragile markets.
First, let us unequivocally cast aside the stifling descriptor, "AI comics." This term, when employed without due nuance, perpetuates a stereotype according to which algorithms can produce meaningful and engaging comics without any human determination, or where human agency is reduced to mere insignificance. Such a description fails to acknowledge the intricate interplay of computational creativity in a medium with a long history of human-machine determination. While speculations about the growing role of automation in artistic production have been a consistent trope in modern and contemporary art debates throughout the mid-twentieth century, comics have always been situated in a dense information economy driven by automation and best practices for transforming craft into mass production. The industrial scale of these operations, from the early stages of ideation to the last-minute editorial revisions, relies on a carefully orchestrated division of human labor and technical determination. Mainstream comics are the direct output of industrial processes of completion based on instituted sets of standardization practices and notions of efficiency, marginal utility and computability that are both conceptually and technically crucial to the comics medium.
Instead of “AI comics,” in Belgium we prefer the term “synthetic comics” (bandes dessinées synthétiques). This acknowledges how comics have always been the products of the aggregate nature of knowledge production and dissemination, increasingly so when they are determinedly embedded in the information-intensive contexts of our computational age. The integration of computational processes for the production of synthetic comics can be traced back to the origins of comics craft and its symbiotic expansion alongside the early development of printing, distribution, and communication technologies. Synthetic comics are works whose forms, material characteristics, and operations are entrusted to technical systems programmed to generate an infinite array of distinctive and original outcomes, leveraging annotated datasets. Their production depends on sets of discrete, decentralized and somewhat asynchronous operations that should be captured in terms of what computer scientist Rudy Rucker describes as computation: any "process that obeys finitely describable rules," involving operations of calculating, processing and transforming information employing diverse substrates, digital or otherwise. Consequently, the significance of computation in their creation cannot be overstated, for it is instrumental in apprehending and constructing the aesthetic encounters facilitated by these works.
I am a comics artist and a postdoctoral researcher working on computational creativity. I have been publishing comic books for over twenty years, using comics less as a vehicle of personal expression and more as a tool for disseminating new forms of knowledge production. Comics is a medium for tackling larger ideas in politics, society and ethics, allowing the production of works that, in other sectors of contemporary research and practice, would be too expensive, too risky, very slow or simply highly dependent on institutional support to initiate.
For myself and many in my Belgian community, comics represent a rich field of experimentation. I've actively delved into this dynamic medium by creating Shapereader comics tailored for visually impaired users/readers, undertaking large-scale, global crowd-sourced projects like Peanuts Minus Schulz or The Cubicle Island, and exploring the capacity of comics objects to articulate artistic statements, elaborate reflections on the medium’s materiality, and establish bridges with the literary arts and conceptual poetry. Comics have also been an opportunity to explore machinic assemblages with books such as Out Side, a collaboration with K Allado-McDowell, founder of Artists + Machine Intelligence (AMI) program at Google—, Fastwalkers, possibly the first comic book co-created with emerging AI—, or the Neural Yorker, a generator of editorial cartoons based on customized generative models. Recently, I mainly focus on designing tools for relational vector databases holding millions of comics media, as well as developing generative engines and computational systems for producing synthetic comics.
Synthetic comics are not transparent, neutral entities. Their computational systems are inexorably rooted in political, social and economic worlds and, more often than not, engineered in ways that reinforce and amplify existing hierarchies— a concept explored by theorists like Donna Haraway, who referred to this as the "informatics of domination." Access to computational tools and platforms is undeniably restricted by socio-economic, geographical, and cultural barriers, resulting in a digital divide where only certain individuals or entities can fully participate and benefit. While I am enthusiastic about the new prospects in synthetic comics, working with computational tools is daunting. Fastwalkers, for instance, required extensive model development to train conditional domain-specific models on the features we needed for our comic narrative, such as characters, landscapes, vehicles, weapons, etc. The production costs accounted for thousands of GPU hours on expensive A100 computational clusters that Nvidia's Inception Program generously provided. As importantly, data extraction processes frequently rely on unrecognized and uncompensated contributions from diverse populations, raising ethical concerns about consent and ownership. The questions of who has access to computational resources, generative models and ample amounts of annotated data, and who provides the data for these models is an important question, and there are ongoing efforts to mitigate the operations of data extraction.
While computational systems often reinforce existing socio-economic and cultural hierarchies, comics artists, have a long legacy in experimenting with and contributing to the expansion of a diverse range of tools ranging from the very early photoengraving technologies, such as the industrially patented Ben Day, to today’s generalized use of proprietary and open source software for image editing, lettering, colouring and layout. Computer-generated imagery (CGI), 3D modeling, algorithmic text generation, and other technologies common in industries such as animation, movie production, and video games have been regularly used by comics artists. Fueled by the inventive spirit and open-mindedness of both historical and contemporary practitioners, experimental comics artists have consistently leveraged the technologies of their time to expand their expressive tools, disseminate new artistic forms, and explore unconventional processes, often with innovative ambitions, such as in Giovanotti Mondani Meccanici's Computer Comics, pioneering computer-generated comics in the 80s. Recognizing the potential of machine learning tools in this context feels more like a somewhat expected progression rather than a disruption, aligning with the tradition of innovation that has long defined the history of comics.
Yet, the apprehension surrounding the introduction of ML tools in comics contradicts the medium’s longstanding, symbiotic relation to technology. I believe it’s fair to write that in a medium that has historically served as a laboratory for today’s global entertainment industries, the vast majority of contemporary comics artists are not only unwilling to see the potential of computational tools, but are also vehemently opposed to the idea of integrating them for the production of their work. Seasoned professionals publicly “disavow AI” and call for a general blackballing of artists who would “deceptively” create work using generative AI. In other cases, publishers called out these “fake” artists who don’t even have “a Facebook or Instagram page” (or what is humorously known as the boomer Turing test) and retailers outright refused to stock comics featuring AI-generated cover art. Additionally, academics still face significantly increased challenges in publishing research related to synthetic comics.
This tension highlights a profound paradox within the comics world. It is surprising that this industry, which has historically embraced "swiping"— the deliberate replication of panels or pages from prior works and arguably one of comics' most innovative contributions to the history of creativity— would now rigidly assert the "human hand" as the only warrant of authentic creativity. While comics communities have long celebrated artistic borrowing and reinterpretation as sources of innovation, they pain to understand how new technologies can help them, much like previous generations refused to acknowledge how digital tools would empower them. Today, few would claim that contemporary comics are a weaker form of expression due to digital photo editing tools such as Photoshop or Procreate.
Some professionals go the extra mile by proposing different variations of industry-wide compliance with a "Made by Humans" label, displaying a nostalgic yet uneasy yearning of the Comics Code Authority's era. As a reminder, CCA was a regulatory body coveted as a third-party alternative to government censorship administrated by the comics industry. It was legitimized by controversial figures like Fredric Wertham, who was known for pathologizing readers and authors, aiming to uphold a "code of ethics and standards" against perceived social misconduct. Setting the question aside of what the political aspirations of mainstream and alternative comics artists who reminisce of postwar US (and there are many!), the tribal, conservative overtones of anti-AI feelings have an evident and dark pedigree in comics history— comic book bans are the forerunners of their twenty-first-century variants of blackballing, deplatforming or doxxing.
There's a noticeable cognitive dissonance between comics professionals decrying AI's impact on their industry and their longstanding use of social media to promote their work. Many outspoken critics of AI within the comics community express concerns about generative AI technologies, yet they conspicuously overlook the broader implications of platform capitalism, which thrives on data extraction. For years, many comics professionals have not just willingly shared content—comics, illustrations, Instagram reels— but their entire networked life on platforms using data to train algorithms and enhance recommendation systems. This disconnect highlights a potential hypocrisy, as these same vocal professionals rarely objected to the "zero compensation" model of sharing their work on platforms like DeviantArt or ArtStation. They also did not scrutinize social media platforms where they've "lent their reputation," nor did they see an urgency to thoroughly review contractual agreements or terms of service. Nobody cared until platform capitalism knocked on their door. As much as this is symptomatic of a "Not Against My Business or Industry (NAMBI)" mindset, this reflection might encourage a more critical examination of all sharing practices, prompting comics professionals to adopt better habits, such as practicing due diligence when engaging with platforms—a potentially positive outcome in an otherwise conflicted stance.
Among the small group of comics artists who have embraced computational tools like Midjourney, an intriguing pattern emerges. Many seem to downplay the creative aspect of their work with these technologies, either minimizing the merit of their AI-assisted creations or hesitantly defending the creative effort involved in crafting effective prompts—acknowledging the discipline and numerous iterations required. This attitude reveals an entertaining hesitance to wholeheartedly affirm their work against the industry's staunch traditionalists and their prevailing attitudes within the community. And yet, neither the act of prompting nor the traditional production of comics is inherently creative as both require a blend of conceptual thought and technical skill. But in comics, the perceived value lies in the tangible, labor-intensive aspects. What saddens me is that despite the increasingly multiple ways one can be a comics artist today, the notion of "hard work" remains the primary measure of an artist's legitimacy and worth within the comics community. This emphasis on effort over innovation or experimentation with technology reflects deeper cultural values about artistry, where sweat equity is seen as the core pillar for being a meaningful member of this community.
When artists decide to take generative AI “seriously,” it’s usually to reveal the concealed, repressed, disavowed and deceptive motivations behind an artwork. Isn’t it twisted that one of the main ways to spot AI-generated art in the commercial world of fantasy illustration and superhero comics, a bastion of male gaze and female objectification, is to look for illustrations of women with sexualized features, beautiful poses and suggestive expressions? I can’t help but wonder why Midjourney excels in that kind of “cheesecake images”, but joking aside, there is an interesting collateral effect here. We observe the gradual institution of an entirely new practice of “suspicious reading” promising to reveal an image’s production context. Comics professionals and anti-AI evangelists are becoming experts in the science of "comics forensics," designed to interpret clues, establish connections, and identify culprits by determining if AI has created a particular image. Besides the proverbial but short-lived seven-finger Midjourney hands, possibly one of big tech’s most fantastic subliminal stunts, an array of subtle tell-tale signs of “dishonesty” are reflected on the reader’s multilayered interplay of affect and expectation: Portraits are dissected for extra creases around the chest, unaccounted human geometries, skin blemishes, facial “distortions”, unexpected body hair and other signs of AI “doctoring". The response to these AI-created images thus serves as a mirror reflecting the community's deeper assumptions and unease regarding authenticity, normalcy, and the evolving role of technology in reimagining human aesthetics. These considerations reveal more of the comics community's normative preconceptions about the human body than anything else.
It's intriguing to consider whether comics artists, who have long championed a genuine vernacular form of artistic expression and often opposed institutionalized art and traditional academia, now are the unexpected defenders of practices centered on mastery. Ironically, comics artists, often marginalized by the art world for their commercial associations and seen as craftspeople rather than artists, now find themselves inadvertently caught in the crosshairs of rampant democratization—not just from generative AI but also from novel forms of media literacy. This role reversal presents a curious twist: comics artists are now the gatekeepers of the established norms they once contested. They are the ones who once fought against elitist art world conventions but now mirror those exclusionary tactics, critiquing new artists for not engaging in the "serious work" they once eschewed.
From memes and rage comics to the emerging skill of promptcraft and synthetic comics, these mediums not only undeniably build on the creative ethos rooted in comics; they also evolve it to meet the aspirations of the twenty-first century. As the boundaries of creativity expand and evolve, this raises questions about how these shifts redefine artistic legitimacy and who is empowered or marginalized in this ongoing transformation. It's no coincidence that the loudest critics of text-to-image generation hail from countries with a rich cultural heritage in comics. Now, it's their chance to admonish artists for not putting in the "serious work" required to make a mark in the comics world.
Synthetic comics and computational tools have the potential to play out multiple infrastructural interests at the crossroads of institutional, educational and commercial establishments. Whether these interests might be integrated into broader strategies with far-ranging impacts on the comics industry is tied to how we understand these new tools and the production of new insights in comics and what futures we can construct from these insights. Understanding how comic artists can harness computational tools to produce synthetic comics contributes to a dynamic reconfiguration of craft traditions. It represents an excellent opportunity to expand the operational agency on the part of comics professionals.
Comics artists, by learning to navigate vast quantities of data, are able to identify patterns and outliers, experiment with innovative data visualization techniques, bring their experience in compelling storytelling for data mining, or rely less on ready-made solutions and straightforwardly build their own customized generative models in the production of image, text and multimodal content. These new skills might reinscribe and reconfigure what it means to be a comics artist in a highly technologically-mediated landscape, what counts as creativity, and possibly help us understand better how to respond to ethical questions related to authorship, attribution and co-ownership.
Comics artists were never just making images. They have been making and contributing to building systems.
Comics artists have approached their craft in social, evolutionary, and dynamic terms to embrace a diverse range of inscription techniques, skills, and expertise that move beyond traditional types of labor associated with making a comic book.
The integration of computational tools in comics is part of a continuous process of technological world-making that is deeply inscribed in the history of the comics medium, intensely more so in countries such as in the US where comics are an essential part of its popular cultural heritage and where comics communities have a collective stake in what the futures of comics would look like.
The research for this article is part of the .expub project (Creative Europe 2023-2025), a consortium between Echo Chamber (Belgium), Aksioma (Slovenia), the Institute of Network Cultures (the Netherlands) and Nero Editions (Italy). .expub explores an operational model for expanded editions and experimentation with new forms of publishing.
The post In Defense of Synthetic Comics appeared first on The Comics Journal.
Hello, my name is Jack Sparrow. I'm a 50 year old self-employed Pirate from the Caribbean.
No comments:
Post a Comment