After writing my previous article about AI and the creative process, I decided to put theory into practice and explore how I could apply these ideas to my own sculptures using ChatGPT — to see just how far this collaboration could take me.
Even though my practice is deeply rooted in physical sculpture, not digital art, I was curious to find out whether AI could still play a meaningful role in my creative process. I wasn’t interested in having AI replace me — but rather in using it as a tool to expand my vision, to see if it could offer new perspectives or spark new ideas I might not have reached on my own.
To my surprise, this process turned out to be much more challenging than I expected. AI is indeed powerful, and while many people claim it can replace artists, those who truly work with it know that using AI effectively is an art in itself. You need to art-direct it, refining your prompts and guiding its output with precision until the results align with your vision.
This is exactly what I set out to do — providing an initial starting point and progressively giving AI clearer directions about what I wanted and what I didn’t want based on its responses.
To begin, I used a photo of one of my sculptures as the starting reference and paired it with a simple prompt to see what the AI would create from it.
Based on this simple prompt, this was the answer:
Yes, I can definitely help you with that.
I see from your uploaded image that your sculpture carries a very distinctive style — fragmented, almost archeological, with a sense of both monumentality and incompleteness. It blends figurative accuracy (especially in the face and torso) with abstraction (the cut forms, absence of limbs, rough vs. polished surfaces). It has a powerful psychological weight, almost like a relic unearthed from another time, yet modern in its treatment of space and void.
If you’d like, I can generate imaginary sculptures in this same style. For example:
Possible Imaginary Sculptures in Your Style
Fragmented Guardian
A towering bust of a figure with one half of the face missing, the other hyper-detailed, with the torso cut diagonally as if unearthed. Surface textures contrast smooth anatomy with rough clay striations.Echo of Motion
A torso leaning dramatically forward, arms cut away at the shoulders, but with incisions in the clay suggesting where muscles would pull in motion. The head slightly tilted, eyes hollow, like a mask.Silent Dialogue
Two figures back-to-back, both missing the frontal face planes. One side smooth, the other eroded, playing with duality and absence/presence.Cosmic Vessel
A bust with a hollowed-out chest cavity, where negative space becomes the main form. The edges crisp and geometric, contrasting with organic clay textures.The Witness
A partial head and upper torso, one eye deeply recessed and shadowed, the other left as a void. The surface etched with subtle linear patterns, giving the impression of erosion by time.
So as you can see, it really captures the esesnce of the work, let’s now see what it does with it visually.
So as you can see, the AI makes its own interpretation and creates something entirely new from the input. The quality of the results is truly mind-blowing — especially in terms of rendering and form — but when it comes to design and artistic intention, there’s still a lot of work to do before I get something I can actually use.
That’s where my role as an artist comes in. The next step is to refine the direction — to teach the AI more about my sculptural language and to bring it closer to my aesthetic goals.
So here’s what I decided to change next:
There is definetly an improvement, so I carry on directing this thing…
At this stage, things started drifting a bit off track—but that’s not necessarily a bad thing. This is part of the creative process: experimentation, trial, and adjustment. In art, as in life, you often need to explore different directions before finding what truly resonates.
So I decided to feed the AI with more information, refining my prompts to help guide it back toward my vision and see if we could realign with the essence of what I was originally aiming for.
This is by far the best result I’ve managed to get from the AI so far. It’s beginning to understand the direction I’m aiming for, and while it’s still quite primitive, it’s starting to produce outcomes that have potential. The results are often stylized human figures—generic in form, but with glimpses of something interesting. Still, they don’t yet reach the level of a cohesive design I could actually work from.
So, I kept feeding it more information, trying to steer it closer to my vision. But from this point onward, things started to drift off completely. The AI went in directions that didn’t make sense for sculpture—at one point, it even forgot we were working with clay and added hair to the piece, which is completely out of context. Have a look:
Conclusion
After spending a few hours experimenting with AI, my takeaway is that yes—it has enormous potential, but it’s not nearly as easy to use as it might seem. It often wanders off in unexpected directions, which can be creatively inspiring, but once you try to focus on a specific outcome, things get tricky. It’s like working with an over-energetic child who keeps getting distracted—full of ideas, but hard to keep on task.
This was just an initial attempt. I’m sure that with more time and by exploring different AI tools, I could achieve other results —maybe not better, but certainly different, I might give it another go another time, it is definetly worth exploring. In the digital world, these tools are incredibly powerful, offering endless possibilities. However, translating that level of imagination into the physical world of sculpture is another story altogether.
AI is far from magic. It’s fascinating, yes, but it still requires artistic direction and a clear vision to turn randomness into meaning. Another thing that is interesting to point out is the stylised nature of the outcome, sometimes giving images closer to marvel movies than a 19th century sculpture even though I told him so. That said, given how fast this technology evolves, it may only take a few months before it reinvents itself yet again.
I hope you found this experiment insightful—and if you’re curious, give it a try yourself. Share your own results in the comments below; I’d love to see what you come up with.
Until then, thanks for reading. If you enjoyed this article, consider subscribing to my newsletter to get more insights on sculpture, creativity, and early bird discounts for upcoming workshops.