Emergent Forms
AI Conceptual Shifts
An exploration into the evolution of AI conceptualization begins with simple prompts—such as “a photorealistic pencil drawing of a hand grenade on a pristine white background”—which generate an initial set of AI-produced images. One image is randomly selected as the seed for the next iteration. The AI then regenerates this image, producing new variations. This process repeats, with each cycle drifting further from the original concept.
As the AI establishes new baselines with each generation, this gradual drift—known as model collapse—mirrors our collective journey as we become more dependent on technology. We now use AI to form ideas about how we see and understand the world, seeking advice on connecting with others and approaching problems both technically and emotionally. By outsourcing more of our cognitive and emotional processes to machines, we introduce significant and unavoidable risks of misalignment and loss of authentic understanding. The AI’s conceptual shifts may not align with human values or truths, transforming our perceptions and realities and often leading us away from natural forms and authentic experiences.
An intriguing instance occurs when an image of planet Earth begins to morph into a solar eclipse. Midway through this transformation, the visual momentarily attempts to form an image of a human fetus—a subject the AI is programmed to avoid. This fleeting moment underscores the tension between technological autonomy and ethical boundaries, raising questions about the latent capabilities of AI and the consequences of relying too heavily on technology to navigate our world.
Example Prompt
“A photorealistic pencil drawing of a hand grenade on a white background”
Algorythm
Generate 4 versions of target
Select a variation at random
Regenerate using selection
Repeat