In Conversation: Andy DiLallo, Jacob Riddle and Ian Anderson on Unplayable AI systems
- Amy Jiang
- Oct 14
- 7 min read
In the feedback loops of machine learning, certain bodies, cultures, and ways of knowing are made visible, distorted, or erased. Like ghosts—excluded yet ever-present—they haunt the machine. They slip through algorithmic blind spots, bend systems’ logic, and disrupt with glitches and ruptures that unsettle computational control. But ghosts don’t just haunt—they rewire. They assert presence, reframe meaning, and speak in forms the system was never trained to understand. Through cracks in the loop, they reemerge as designers of new imaginaries for themselves and for the communities the system failed to see. Ghosts in the Feedback Loop is a virtual exhibition that invites artists to work inside those cracks—to treat algorithmic systems not as endpoints of automation but as haunted infrastructures alive with memory, loss, and rebellion. This interview is part of TechnoMirage, UAAD’s latest curatorial & publishing project exploring the intersections of artificial intelligence, speculative design, and collective imagination. Emerging from a multi-format event series—including a virtual exhibition, an online panel co-hosted with Parsons, and an IRL gathering of workshops, talks, and performances—the publication extends these dialogues into an archival form.
About the Artists
Andy DiLallo is a new media artist whose work explores the ideological underpinnings of algorithmic systems, machine learning models, and digital agents. His practice examines how computational protocols shape emotional experience, desire, and agency, often critiquing the incentive structures of platform capitalism while imagining alternative logics. Andy has exhibited internationally at venues including SIGGRAPH, Athens Digital Arts Festival, IndieCade, The Wrong Biennale, and COP26.
Jacob Riddle is an interdisciplinary artist, curator, and educator whose practice bridges technology and the natural world, informed by his background in construction, labor industries, and his Appalachian upbringing. His work emphasizes resilience, resourcefulness, play, and harmony between digital and natural environments. Jacob has completed artist residencies in Finland, Norway, Berlin, and Nez Perce-Clearwater National Forest, with exhibitions spanning the United States, Mexico, Japan, Europe, Australia, and various virtual platforms.
Ian Anderson is a professional 3D modeler, animator, and developer, most often working collaboratively with physicians in his role at Cincinnati Children’s Hospital. He produces 3D animations and virtual games designed to mitigate pain in children, educate nurses, and conduct VR research. His practice translates expertise in 3D animation, video production, and rendering to explore the visual, sonic, and world building possibilities of virtual space.
Q: Could you tell us a bit about your background, and how your work connects to the theme “Ghosts in the Feedback Loop”?
We first began working together as part of the Utopian Megaproject, a non-situated collective of artists, curators, and thinkers speculating on equitable futures across time zones and disciplines. At Tryst Art Fair in Los Angeles, we prototyped our collaboration by combining our respective practices into a shared framework rooted in exploring error, failure, and the productive potential of loss.
That ethos carries directly into Unplayable. One of our early prompts was Hayao Miyazaki’s claim that AI cannot understand pain and therefore cannot make art. We became interested in how the machine’s incapacity might reflect back on us: our failures, ruptures, and blind spots as humans are not deficits but precisely what differentiate us from systems designed for efficiency and optimization. Within this frame, the “ghosts in the feedback loop” are the moments of unknowing, the irresolvable errors, the spaces that slip outside the binary logics of computation. These hauntings remind us that what makes us human often emerges in the very places machines cannot reach.
Q: How do you perceive AI—as a tool, a collaborator, a medium, a subject, or something else? And how does that shape your artistic process?
We approach AI less as a tool or collaborator and more as a recombinant system inseparable from the data it is trained on. In many ways, it is a mirror of human language, an immense archive of our cultural expressions, biases, and contradictions, recombined into outputs that are coherent but never embodied.
This is where psychoanalytic theory becomes generative for us. Lacan’s notion of the objet petit a, the unattainable object-cause of desire, marks the limit of language. For AI, such a concept is like a paradoxical command, a kind of Roko’s basilisk: it cannot resolve that which exceeds symbolic capture. In this way, AI exposes its own boundary as a disembodied language system.
While AI is built on vast amounts of quantitative data derived from human traces such as our clicks, words, and images, it cannot reach the non-quantifiable aspects of human life: pain, intuition, care, and contradiction. These residues, which escape metrics, are precisely what differentiate us from systems designed to optimize and predict.
Artistically, this opens a powerful mirror stage. When we engage with AI, we encounter both a reflection of our cultural data and an index of what escapes it, what cannot be reduced to recombinant language. Our process leans into this gap, using AI not to simulate creativity but to illuminate the fault lines where human knowledge, embodied experience, and failure exceed the system.
Q: What futures does your featured work gesture toward or warn against? Who do you imagine as your audience, and how do you hope they are impacted?
Unplayable warns against the erosion of agency within platform dynamics that compress our lived experience into loops of “platform time,” including feeds, infinite scrolls, and algorithmic cycles. In this world, memory and difference flatten into repetition, and the possibility of meaningful interruption dwindles.
Against this backdrop, we imagine two paths: to become “non-playable,” unreflectively reproducing the logics of the system, or to become “unplayable,” a refusal of those logics through withdrawal, disruption, or refusal to perform as the platform demands. The work urges literacy in recognizing how these systems shape our sense of self and agency in resisting them.
At the same time, Unplayable gestures toward futures of solidarity. No one can step outside these dynamics alone; it is only through community that resistance gains weight. In this sense, our imagined audience is less an individual viewer and more a collective of witnesses. We hope the work unsettles their position just enough to make them reflect not simply on how they resist alone, but on what solidarities we can build together, even as NPCs in a system stacked against us.
Q: Are there particular communities, histories, or environments your work remains in conversation with? How do those relationships evolve over time?
Our work is deeply informed by online communities that foreground discourse around platforms, media, and technology. The New Models community, particularly Caroline Busta and Lil Internet’s articulation of “Platform Physics,” has provided a language to think about how infrastructures shape culture at the level of perception and behavior. Similarly, spaces like Do Not Research reading groups and Trust in Berlin sustain dialogues that are less about fixed political programs and more about experimenting with forms of collective inquiry in the wake of eroded public space, especially in the United States.
We are also in conversation with artists and researchers who are actively prototyping new cultural infrastructures for AI. The work of Mat Dryhurst and Holly Herndon, particularly the recent Choral Data Trust Experiment, demonstrates what it might mean to treat AI training data as a public good stewarded by communities rather than monopolized by platforms. These experiments with governance, stewardship, and collective rights resonate strongly with our interest in how cultural institutions can reimagine their role in shaping technological futures.
Taken together, these communities and projects remind us that while platforms flatten and privatize discourse, parallel infrastructures of care and inquiry continue to emerge. Our work exists within that ecosystem, both haunted by the platforms we critique and buoyed by the communities imagining alternatives.
Q: What do you see as the most urgent threats or uncertainties we may face in the coming decade with the rise of AI?
We see the most urgent threat less in AI itself than in its convergence with what Neil Postman diagnosed in Amusing Ourselves to Death: the erosion of public discourse through entertainment. Already we see the phenomenon of “sycophancy,” in which models tell users exactly what they want to hear, even at the expense of truth. When content is personalized to each individual, our common ground fragments and spaces for collective dialogue collapse.
At the same time, AI is inseparable from the political economy of data. Models are only as good as their training sets, and those sets are the product of extraction, enclosure, and ownership struggles. In this sense, AI extends the logic of platforms, harvesting user activity as currency while degrading the very quality of the systems we rely on through what Cory Doctorow has called “enshittification.”
The challenge is therefore both ontological and political. Ontologically, AI reshapes how meaning and communication are mediated, increasingly blurring the line between fact, fiction, and entertainment. Politically, a handful of corporate actors control the infrastructures of data and attention in a way that resembles the robber barons of the industrial age. The uncertainty is whether we can build counter-infrastructures, forms of stewardship, commons, or public-interest AI before those monopolies solidify beyond challenge.
Q: Are there any theories, books, or artists you’d like to recommend in your current areas of interest?
Yes, our thinking has been shaped by a constellation of theorists, artists, and communities who grapple with subjectivity, platforms, and simulation.
We are indebted to Marek Poliks and Roberto Alonso Trillo’s recent book Exocapitalism, which pushes discourse on labor and production into new terrain, accounting for everything from high-frequency trading to NPC culture in online games. Their earlier workshop text Non-Player Dynamics has been foundational to our exploration of agency and its limits in game worlds.
We would also point to Jean Baudrillard’s Simulacra and Simulation, which feels newly urgent in an era where sycophantic fictions blur seamlessly with political reality, and where Elon Musk’s claim that “the most entertaining outcome is the most likely” echoes Postman’s warning about entertainment as the dominant mode of control.
Artists and researchers like Holly Herndon and Mat Dryhurst, particularly their work on the Choral Data Trust Experiment, exemplify how artistic practice can intersect with infrastructural invention. Likewise, the New Models community continues to generate indispensable frameworks for understanding platform dynamics.
Lead Editor: Amy Xiaofan Jiang
Assistant Editor: Paridhi Garg