CAN-not

CAN-not is part of the ongoing project CAN-verse, exploring standardization, control, and transformation within mass systems through the image of the aluminum can, a symbol of industrial production and consumption.

The work takes the form of a three-channel AI video installation composed as a single triptych. It consists of three vertical screens, each divided into four horizontal sections resembling both store shelves and conveyor systems. All three videos begin from the same source image.

Each screen is generated using only one text prompt defining the behavior of the system. In total, the installation uses three prompts. The first establishes a highly controlled and predictable model of movement. The second reduces the level of control, allowing unstable and unpredictable interactions to emerge. The third gives the generative system near-complete freedom to transform beyond its original structure.

Each prompt is repeatedly processed through a generative AI system without manual intervention or correction by the artist. Despite sharing the same initial conditions, the three systems gradually diverge, producing different trajectories, structures, and forms of behavior over time.