Conceptual framing
Generativity troubles authorship. From Vera Molnár’s algorithmic drawings to AI image synthesis, artistic intent is always entangled with algorithmic unpredictability. Who generates? And who owns the work?
Description
This module introduces co-authorship with AI through iterative workflows in ComfyUI. Students experiment with image generation, variation, and remix, while exploring dataset curation as an aesthetic and ethical decision. Outputs from TouchDesigner generative sketches are looped into AI pipelines, producing hybrid works. Through these exercises, participants encounter AI not as tool or autonomous creator, but as a collaborator that introduces divergence and surprise. The module frames generative practice as a negotiation across human imagination, machine agency, and data.
| Public sessions (open to all) | Workshops (sign-up required) | ||
|---|---|---|---|
| 11 Dec (Thu), 7pm | Introductory lecture + artist talk | 13 Dec (Sat), 10am | Workshop 1 |
| 20 Dec (Sat), 10am | Workshop 2 |
Workshop 1 introduces participants to generative design through TouchDesigner and establishes a foundational workflow for integrating ComfyUI as an image-to-image processing pipeline. Participants will learn the basics of node-based construction, procedural pattern-building, and exporting imagery from TouchDesigner before using ComfyUI to transform these outputs through AI-driven variation. The session culminates in a hands-on exercise in which each participant creates a series of generative flower forms—first designed in TouchDesigner, then reinterpreted through img2img in ComfyUI—to reveal how human intent and machine behaviour intersect in iterative visual processes.
Hardware
Software
Latest stable version:https://derivative.ca/download
ComfyUI Portable (Win64)
**https://github.com/comfyanonymous/ComfyUI/releases**
Model Requirements