Memoranda of Artificial Understanding is a collaborative-responsive project by Brandon Gellis and Sue Beyer that explores the unstable boundary between instruction, interpretation, and invention. Working with AI-enabled content generation, the project embraces both system and accident, using rule-based prompts, generated outputs, and hallucinated imagery to reveal how meaning emerges as much through misreading as through intention.
Rather than treating AI as a neutral tool, the project approaches it as a co-producing agent whose outputs are shaped by constraints, drift, and unexpected invention. The work is interested in what happens when artistic rules are precise, but their results remain unpredictable, and when machine logic produces visual or textual forms that feel at once legible and estranged. In this sense, the project becomes a memo—not a fixed statement, but a record of human-machine exchange as it unfolds.
At the center of the project is a question about authorship, reliability, and the aesthetics of error. AI hallucination is not framed here as failure alone, but as a generative condition that exposes the instability of image-making and knowledge itself. By combining rule-based structure with responsive collaboration, Memoranda of Artificial Understanding proposes that artistic meaning can arise from partial control, recursive feedback, and the productive uncertainty of shared authorship.