Moth Glitch

Computer-generated moths struggle to flutter against a backdrop of glitched, failed AI outputs. These colorful patterns emerge from a paradox: asking a diffusion model, which operates by stripping noise from a jpg toward an image described by a prompt, to produce an image of the noise it is designed to remove. The result creates a glitch in the system, superimposed with stilted, toylike moths – the original “computer bug.” Moth Glitch is an ode to Stan Brakhage’s Mothlight, where real wings served as a replacement for photographs of their wings on film. Moth Glitch attempts to treat glitched images as an artifact, like spoiled film. The dance of the moths suggests a joyful confrontation with the boundaries of systems that retrieve images using language, as the soundtrack declares: “shaped by language / words assigned to an experience / shape what returns to us.”

Eryk Salvaggio
Eryk Salvaggio (cyberneticforests.com) works critically with AI-generated media, using and subverting tools to reveal their embedded logics. While working critically with technology, his work acknowledges the paradox of complicity while aiming for a new visual vocabulary: glitching these machines, repurposing generated imagery to reveal biases or celebrate “errors” in a practice that examines the image’s relationship to its social, technical, and cultural context. Through video essays, sound art, and performance lectures, Salvaggio’s work confronts essential debates in AI while such as the politics of extraction, dehumanization, and information pollution that lay at their heart – while acknowledging that other AI’s are possible. “Many artists see AI as a tool for expanding the imagination,” Salvaggio says, “But I aim to pull the imagination out of AI, to see it for what it is, so we can make more thoughtful decisions about our creative relationship to it.”