Carlos Fentanes

One afternoon, while leafing through a listless scroll of open calls—those desperate pleas for money masquerading as Art opportunity—I came across something peculiar. It wasn’t flashy, nor particularly detailed. Just a line or two, almost buried among more demanding announcements, It simply asked for artwork generated by artificial intelligence.

“Seeking artworks  generated with AI ( the acronym for Artificial Intelligence). Theme: Future: Beyond Imagination.”

That was it. No boundaries, just the requirement that the work be born from a machine. It read less like a prompt and more like a riddle, as if it had been placed there not by a curator but by something else—something watching.

There was no mention of form or function. No stylistic preference. Just a machine-born origin, as if that alone were enough to call it art. I reread the message several times. It began to haunt me—not for what it asked, but for what it didn’t. Most submissions, I assumed, would be images spun from Midjourney or DALL·E—those dreamy, oversaturated AI visions we’ve all seen. Impressive, sure, but more like visual fast food: quickly consumed, quickly forgotten. I’m not the kind of person who enjoys feeding DALL·E prompts, but I do spend a lot of time with ChatGPT—writing code, editing my blog, translating documents.

That’s when I decided not to play along directly, but to respond obliquely. I will ask ChatGPT to generate a code that can create an ever evolving image, so if you come tomorrow you will see a different image of what you see today, but with memory, you can recognize the traces from the previous image on it.

It’s like playing the game of AI without the generative image from AI, swimming against the current to say so. The idea took root in a familiar place. I remembered Marcel Duchamp’s urinal—Fountain—signed with the now-legendary pseudonym “R. Mutt.” His provocation wasn’t the object, but the act. His message: art is not the thing itself, but the thought behind it. I wanted to do the same—make a gesture that blurred authorship, process, and time.

So I turned to ChatGPT. We began constructing a script—a simple one, at first. Each day, the system would generate a single colored dot and place it on a screen. Over time, as the dots accumulated, a picture would emerge—not pre-designed, not fixed, but growing, shifting, forming slowly like memory itself.

The work became an evolving record of “decisions” made by chance, coded randomness layered with time. To me, this is the future: not a sci-fi skyline, but the quiet, relentless layering of moments. And like life, it left a trace. Nothing is erased. Every dot matters, even the ones that seemed misplaced. Especially those.

But the idea didn’t stop there. Once the first version was submitted, my mind kept working. I began experimenting with constraint. What if I introduced text—say, the word Hello—and used it as a filter? The script would only allow color dots to appear on the black pixels of the word, never the white. Suddenly, meaning emerged through absence. Over days, the word would slowly reveal itself, like a message sent into the future.

I tried it again, this time with grayscale images—subtle portraits, tonal gradients—translated into landscapes of color through controlled randomness. I posted these experiments to Instagram and YouTube. The response was modest, but curious. People didn’t just see images; they watched them form.

Somewhere in the process, I was reminded of my own pen drawings—meticulously constructed from tiny dots, built over hours and days. There too, the mark-making is mechanical, repetitive, almost algorithmic and with some kind of randomness. But it’s tied to my hand and my authorship. So how different is my AI-generated portrait, really? It may be made by code, an AI generated code but guided through me, who is the author there? Me or the thousands pages of information that AI had to learn to generate a code according to my orders? The machine is just a tool like a brush or an oil color, we just need to learn how to use it.