Looking forward to being a CAA Co-Chair for SAC roundtable The Politics of AI as a Medium of Art with Lorelei d'Andriole, Michigan State University and Julian Haladyn, OCAD
“Machines have the morality of their inventors.” -Amiri Baraka
How is our current age of technological reproducibility unique from the crisis of the modern era? Ninety years ago, Walter Benjamin wrote “The Work of Art in the Age of Its Technological Reproducibility” introducing political art theory on how technology developed under capitalism shifted the value of art. This roundtable seeks to bring together scholars and artists to discuss the conceptual implications of AI as a medium of art as well as common contradictions of the medium today. For example, the use of computers by new media artists in the United States to produce their work is explicitly intertwined with the violent exploitation of the global south and thus, using such a medium to make work about equality may be complicated to say the least. This panel will not be focused on what AI as a medium of art may do in the future, but is instead focused on what it is doing now and what relevant historic contexts led us to our current crisis within politics and art. Regardless of how artists might use AI in their work we must hold the dialectic that whatever accessibility features or delight it offers, that machines have the morality of their inventors. In this round table, the ideology of the corporate inventors of these tools will be considered as a prerequisite to any individuals artistic expression.
Conjuring Queer Poltergeists: Queer Gothic AI and the Erotic Image
Morris Fox, Concordia University
This presentation proposes a queer gothic engagement with AI as a medium of conjuring, where images arrive as both spectacle and haunting. Just as Victorian spiritualist photography staged ectoplasmic presences through sleight-of-hand, AI generates its images through machinic presage: a conjurer’s act of distraction that obscures AI's ecological costs. Data centers consume water and energy, transforming desire into depletion, yet the spectral images AI produces, especially in erotic and pornographic registers, circulate as if immaterial. I examine how AI art is haunted not only by infrastructures of extraction that leave ruins in their wake, but also by its promise of queer worldbuilding; from Grindr’s AI “wingbuddy” to OnlyFans’ proliferation of synthetic bodies.
Drawing on Billy-Ray Belcourt’s Poltergeist Manifesto, José Esteban Muñoz’s notion of “ephemeral evidence,” and Eliza Steinbock’s theorization of “shimmering images” as queer spectral aesthetics, I frame AI-generated desire images as libidinal hauntings enacting a doubled erasure: an ungraspable desire-image sustained on degrading the very worlds they emerge from. I exemplify this in120 Days of Goosebumps, a horror-porn slideshow I created using prompt-based AI, and in dialogue with Em X. Liu’s queer Hamlet adaptation The Death I Gave Him, where Horatio is reimagined as an AI companion, I ask how ghostly intimacies with machine “intelligence” eroticize and occlude political aesthetics.
To engage AI as a spectral medium is to wrestle with its necropastoral aesthetics: it conjures pleasures and ephemeral communities even as it devours energy, water, and lifeworlds into ghostly debris, flattening the queer horizon it promises to extend.
Beyond the Witch Hunt: Teaching AI, Creativity, and Ethics
Chanhee Choi, University at Albany
As an artist and educator, I first engaged my students in a course on AI, creativity, and ethics while teaching in the Film and Digital Arts program at the University. At that moment, the Hollywood writers’ and actors’ strikes had intensified public distrust of AI-generated works, and my class became a flashpoint of suspicion. Despite my intention to foster critical discussion, I faced accusations that reflected more of a cultural “witch hunt” than a fair critique of pedagogy.
This experience has since sharpened my understanding of the contradictions embedded in AI as an artistic medium. On one hand, AI can expand accessibility, generate new creative languages, and transform collaborative processes. On the other, it raises urgent concerns around labor exploitation, authorship, and the ethical limits of machine-made expression. As artists and educators, we must ask how we can engage these tools responsibly without erasing the human stories and power structures that shape them.
In this roundtable, I will reflect on how to navigate this contested landscape both in the classroom and in one’s own practice without falling into unproductive cycles of fear or blind embrace. I will share strategies for positioning oneself as an artist and educator in an era of technological mistrust, cultivating dialogue rather than division, modeling critical engagement for students, and resisting the cultural impulse toward “witch hunts” that stifle nuanced discussion.
Sludge Aesthetics: Generative AI, the Sublime, and the Limits of Political Engagement
Abhishek Narula, University of Michigan
I argue that the aesthetics of GenAI art are politically impotent. While there are pressing political concerns surrounding GenAI, such as environmental impact,proliferation of misinformation, deepfakes, and surveillance, my analysis will focus on aesthetics. I contend that the aesthetics of GenAI fall within the category of the sublime and, thus, fail to catalyze any meaningful cultural or politically progressive projects.
To frame aesthetics in relation to creative practice, I will draw on Graham Harman’s Object-Oriented Ontology (OOO), particularly his distinction between the sublime and beauty, which diverges from Kant’s formulation. For Harman, the sublime marks the mind’s inability to comprehend overwhelming or formless phenomena, e.g. the immensity of a tsunami or the vast networks of mycelia beneath the soil. Beauty, in contrast, emerges from the allure and enigmatic qualities of objects that remain partially withdrawn from access.
Building on this distinction, I will introduce Miguel Carvalhais’s methodology for reading computational art. Though generative AI employs computational tools, it differs significantly from computational art. For Carvalhais, computational art fosters the construction of mental models in the viewer, enabling interpretive engagement despite its black-box elements. Generative AI, however, collapses this possibility: its hyper-abstraction eliminates the “box” altogether, preventing viewers from engaging in critical reading of artworks.
Thus, I argue that the aesthetics of GenAI are closer to the sublime aesthetics of nature than to those of computational art. This produces “AI sludge”: a mode of hyper-abstraction that resists critical aesthetic engagement and, ultimately, fails to generate cultural or political transformation.
Ending Slop For Slop’s Sake: GenAI As An Accessible Political Medium
Matthew Magill, Michigan State University
In “The Work of Art in the Age of Its Technological Reproducibility”, Walter Benjamin entertains Karl Marx’s concept of an intertwined Base and Superstructure; specifically its role in class antagonism and artistic expression. Marx argues that the Base, which is the ownership of the means of production and property relations between classes, is always shifting. Since these relationships are inherently political, so is the art created under this system. Benjamin goes on to condemn “art for art’s sake” as a tool that rejects this imperative social function of art. As the use of GenAI becomes more profound in our society, it can be easy to proclaim widespread access to this novel tool as a shift in the Base. In my presentation, I will discuss whether the increased accessibility to GenAI should be considered as a shift in the means of production as described by Marx and Benjamin. On a more specific level, my presentation will focus on how this sudden change comes at an observable cost to the Proletariat class, the upheaval of human value in the artistic process and how this affects the way in which we create art, and how GenAI can be used as a political medium in contrast to the meaningless slop created by those who often overlook the dire political implications of this emerging technology.
Anthropomorphizing the Capitalist Sublime
Brooks Cashbaugh, Cornell College
While seemingly harmless, the discourse surrounding art produced by Artificial Intelligence limits a comprehensive understanding of the technology and in doing so mystifies it, contributing to an ideology that protects AI software as an elite tool. The purpose of this research is to parse how AI art fits into a broader web of technology, culture, and socio-economic power. My methods of research primarily focus on analysis of existing literature. This exploration was initially inspired by my own experiential research using GAN (Generative Adversarial Networks) image generating software.
This presentation analyzes criticism and reviews of GAN art from the late 2010s to identify early trends of anthropomorphizing language in the literature. This is contextualized with Marxist theory to identify a larger pattern of anthropomorphizing rhetoric to describe artificial intelligence in a wider field of discourse. This is connected to the increasingly prominent role that artificial intelligence software plays in our world. My presentation goes on to examine AI art in light of Sienne Ngai’s gimmick theory to explore AI art’s connection to the capitalist sublime. Ultimately, my research finds that while AI art obscures or softens some aspects of broader AI uses, it still functions as a Ngaian gimmick, revealing the technology’s inescapable entanglement with a global system of exploitation and extraction.