The Negotiated Line

Diego Silan

Mini-Essay: The Negotiated Line

Concept and Motivation

In this project, I explore the techniques I applied after our collaboration assignment, where we experimented with misspelled prompts to see how AI interprets flawed language. That experience showed me how powerful prompt writing can be and how AI sometimes corrects or misreads intention. This became the foundation of my individual reflection because it connects directly to both the group investigation and my personal creative inquiry. I wanted to push beyond typing prompts on their own and see how they interact with sketches. My goal was to learn how much influence a person retains when AI becomes a collaborator. Working with sketches helped me clarify this motivation. I wanted to understand whether carefully designed wording could transform a hand-drawn idea into something believable and refined, letting me observe how AI perceives intention rather than simply executing commands.

Hybrid Workflow: Sketch-to-Image Prompting

After the collaboration assignment, I chose sketch-to-image generation in Runway as a method to learn more about prompt literacy. This workflow forced me to communicate visually and verbally at the same time. Drawing simple forms let the AI add detail and depth, transforming my basic idea into something aesthetically richer. It was fascinating watching a preliminary sketch suddenly appear polished. My goal was not to let AI invent everything but to intentionally direct it through well-crafted language. I wrote prompts describing features to include, such as lighting, face shape, or mood. However, this workflow also showed me how easily a prompt fails if wording is vague. When I asked Runway to “generate a realistic version” of my sketch, it produced Angelina Jolie—an output far from what I intended. That challenge made me rethink how AI perceives form.

Thinking Evolution

Once I realized AI misinterprets sketches, my thinking evolved. I refined the prompt, asking AI to match the nose shape, chin structure, and eye form in my drawing. When I added these details, it produced Adrien Brody, whose features aligned more closely with my sketch. This shift taught me that creativity with AI is not passive; it demands iteration, specificity, and revision. The work became less about watching AI perform and more about learning how to communicate my intentions effectively. This experience deepened my understanding of creative collaboration. It taught me that art making with AI is not about telling a model what I want one time—it requires testing, refining, and aligning language until the output reflects what I envisioned. Generating meaningful outcomes became a process of problem-solving, bridging analytical thinking and artistic intuition.

Authorship and Human Responsibility

Throughout the process, I ensured that I—not the AI—made the primary artistic decisions. The sketch came from me, the prompts were my choices, and the revisions reflected my judgment. I did not allow the AI to choose what to draw or what meaning to assign. I treated it as a tool reacting to my guidance. This made me reflect deeply on authorship and responsibility. The AI enhanced the work, but I shaped its direction. Another challenge emerged: AI can make artwork look polished, which could mislead someone to believe the machine is the author. However, maintaining agency required constant decisions—what to include, what to correct, what to reject. This helped me understand human–machine collaboration as negotiation, not automation. Good results depended on my capacity to lead the system thoughtfully.

Course Connections and Project Extension

This project clearly connects to course ideas like prompt literacy and human–machine collaboration. Writing prompts became a language exercise where I had to communicate clearly to make the AI understand my visual intention. The AI contributed technical changes, but I provided purpose. This mirrors our group project, where we asked whether wording affects visual outcomes. My experiment confirmed that it does—prompt phrasing materially shaped how AI transformed my drawings. This also relates directly to my specific contribution to the group project, where I designed conceptual visual metaphors demonstrating how different prompts create different aesthetics. My personal experiment extended that research, letting me test prompt influence firsthand through my sketches. It taught me that AI collaboration only succeeds when humans stay intentional—deciding, revising, and guiding results. Creativity becomes shared, but direction remains human, revealing prompt writing as both a technical skill and an artistic practice.

Final Insight

Through this project, I learned that AI art is not an automated shortcut but a negotiation between intention and interpretation. The process demanded curiosity, revision, and decision-making, proving that meaningful results depend on the human artist’s vision rather than the tool alone. Working with Runway helped me recognize that prompting is not only a technical skill but a creative language that shapes aesthetics, authorship, and collaboration. Most importantly, this experience allowed me to develop my voice as an artist in AI spaces, showing me that the value of the work lies not just in what the machine generates, but in how I direct, refine, and learn from it.

Chat Log + Tools Used

AI Tools Used:

  • RunwayML (Sketch-to-Image / Video Generation)
  • ChatGPT (for reflection drafting and editing)
  • Eleven Labs (for text to speech generation)

Other tools Used:

  • Premiere Pro (for putting all the clips and audio together)

AI Session Links: