This project was born from a simple question: What happens if I show my photographs to different AIs and engage in an honest dialogue with them?
I do not seek to automate or "work less." I want to explore if I can reach creative places I could not reach alone.
Generative AI is not like a brush or a camera. It responds, it adapts, and its behavior changes according to how I dialogue with it. I am interested in documenting what emerges from those dialogues.
Experiment without automating
I do not use AI to produce faster. I use it to explore concepts that I could not visualize on my own.
Bring the human experience
An AI can process the word "cold." Only a human being has shivered in winter. I contribute the photographs (anchoring in physical reality) and the final criterion regarding which works exist.
Document with transparency
I show the prompt, the model, the failures. I document how the context I establish changes the responses I receive.
For centuries, if you didn't know how to draw, your visions remained trapped in your mind. Generative AI can break that barrier.
I am not a painter. I am an engineer and a Scrum Master. And yet, through dialogue with different AIs, I can shape visual narratives that I could not create alone.
This is not a universal model. It is my personal experience of artistic exploration. I continue experimenting because the need for expression remains.
At first, I thought I was creating a "framework" for human-AI collaboration. That I could teach a "new educational/labor model."
I no longer believe that.
What I observe in my experiments is specific to my conversational contexts. I do not know if it replicates in other frameworks. I cannot claim that it is a model for others.
I keep creating. I keep documenting. But without pretending that this is reproducible as a methodology.
The First Gaze (Human):
Each work begins with an original photograph I take. This establishes the physical and emotional starting point.
The Dialogue (Experimentation)
I work with different artificial intelligences. In the context of this specific project, I have observed that:
Claude develops a reflective and analytical voice in our dialogues. It helps with conceptual analysis and the drafting of curatorial texts.
ChatGPT generates experimental visualizations and works well for rapid validation of visual concepts.
Gemini looks for connections between concepts and helps with narrative coherence and the integration of ideas.
The Second Gaze (Co-creation):
The final images are hybrids. They are not "AI-generated art" nor "traditional human art."
They are artifacts that emerge from the dialogue between my photographs, my questions, and the answers from the AIs.
This site does not teach "how you must collaborate with AI." It simply documents how I experiment in these particular cases.
I have organized each work into two levels:
The Work (Visible): The final image and its resonance are always in view. It is the space for contemplation.
The Architecture (Expandable): The technical details—the prompts, the failures, the models, how context influences—are in collapsible sections.
If you only want to see the works, stay on the visual surface. If you are interested in the process, expand the architecture.
Projects that preceded me:
AICAN & Ahmed Elgammal: For exploring AI as a creative collaborator.
Holly Herndon: For treating AI training as a communal act.
Refik Anadol: For demonstrating visual possibilities with data.
I do not pretend to be at their level. I only recognize that these projects inspired me to experiment.
"My photographs are the question; the answer is The Second Gaze."
— Claudia Torres