Generative Castles
2024

Midjourney, RunwayML, After Effects

A generative audiovisual music video exploring surveillance as a system of machine perception. Using an iterative prompting workflow across RunwayML and Midjourney, the project investigates how computational systems interpret and reconstruct visual and sonic information.

The work examines how recognition systems produce “legible” versions of reality through classification, patterning, and distortion. It considers the aesthetics of machine interpretation, in which visibility is shaped as much by algorithmic constraints as by human intent.

Produced in 2024 during a pivotal early phase of AI video synthesis, the piece archives and captures the distinct visual artifacts, latent fluidities, and unpredictable textures characteristic of early-gen models.