Woordworking with AI

From Woodworking to Sculpture

This journey began years ago when I first explored woodworking as an amateur. I started small—bowls, chopping boards, and gradually moved on to furniture, crafting chairs and other pieces.

As an architect, I've always been in contact with art. It's embedded in the nature of the trade; we often think of ourselves as artists, even if we haven't yet discovered our own medium.

Eventually, I found myself craving a new creative challenge. I had exhausted ideas (and space) in my home and decided to try my hand at wood sculpture.

A Voice Yet to Be Found

Sculpture, however, was an entirely new language. Without formal training, my early work felt derivative—echoing existing styles rather than expressing something uniquely mine. Through research, I found that most artists emphasized two essential elements in the creative process:

Feedback from experts – Insight from experienced sculptors helps refine an artist's direction. Relentless iteration – The only way to develop a voice is through continuous creation and refinement. I had no issue with the work itself. But I lacked direct access to the sculpture scene—the network of artists, mentors, and critics who could provide meaningful feedback.

Where Technology Meets Creativity

This is where my other career came in. As a software developer, I am constantly testing the limits of new technology. And at this moment, one of the most powerful and accessible tools available was ChatGPT, a large language model (LLM).

Could I use AI to replicate the feedback loop I needed?

This question led me into an experiment—one that would challenge my assumptions about creativity, learning, and the role of artificial intelligence in art.

Bridging the Gap: Bringing My Sculptures to AI

To organize my journey, I created a dedicated project in ChatGPT to group all my sculpture-related conversations. If I was going to treat AI as a feedback tool, I needed continuity—something that would allow me to track my progress and refine my approach over time.

The first challenge was deciding how to present my work. Explaining a sculpture in words felt like a daunting task. Describing form, texture, movement, and intent through language alone seemed not only exhausting but also prone to misinterpretation. I risked injecting meaning that wasn't truly present in the piece itself.

So, with a "here goes nothing" attitude, I uploaded a video of my sculpture to the chat.

To my surprise, the model accurately identified the object and provided a remarkably precise description. This moment sparked a deeper curiosity about how it was processing visual data—a topic I plan to explore in a future post.

But for now, my focus was on feedback. I asked the model to be as critical as possible, to analyze both the strengths and weaknesses of my work, and to suggest artists whose styles might inspire or challenge me.

What followed was an unexpectedly rich dialogue—one that would push me to rethink not just my sculpture, but my entire creative process.

A New Kind of Critique: AI as an Art Critic

Curious about how well AI could analyze a physical sculpture, I decided to put ChatGPT to the test. Could it offer the kind of meaningful critique that artists seek from peers and mentors?

I started with a simple question:

"If I show you a video or photo of a sculpture, can you give me feedback?"

The model responded:

"Yes, you can upload a video or photo of a sculpture, and I will analyze it to provide detailed feedback on aspects such as composition, craftsmanship, material use, proportions, and overall aesthetic quality. Feel free to share the media whenever you're ready."

Encouraged by its confidence, I uploaded a video of my piece. The sculpture had an organic, rough-hewn aesthetic—a suspended wooden form balanced on a thin metal rod. I wasn't expecting groundbreaking insights, but I was open to anything that could push my work forward.

Here is the video, you can see that while it's clear, it has a lot of visual noise.

Then, ChatGPT delivered its critique.

Blunt, Constructive Feedback

Strengths:

Material Choice: The natural wood texture is visually engaging; the cracks and imperfections contribute to a rustic aesthetic. Floating Effect: The suspended presentation creates tension and intrigue, enhancing visual interest.

Weaknesses:

Form Ambiguity: The piece lacks a clear visual narrative or intentional abstraction, making it feel more like an unfinished block than a sculptural statement. Surface Treatment: The rough-cut marks and natural damage overpower any sense of deliberate craftsmanship. Structural Vulnerability: The thin metal rod creates visual imbalance and might not be structurally sound over time.

Base Integration: The base feels disconnected, competing with the sculpture rather than supporting it conceptually.

Suggestions:

Refine the shape: Either sharpen geometric precision or fully embrace organic fluidity—this middle ground feels weak.

Surface work: Introduce contrasts—smooth vs. rough textures—to create a more intentional finish.

Strengthen the narrative: Right now, it's more "interesting wood" than "art." What's the concept? What emotion or symbolism should it evoke?

Base cohesion: Either refine the base to match the sculpture's language or intentionally lean into its rawness.

The critique was shockingly direct. It didn't hold back, and it pinpointed issues I had been vaguely aware of but hadn't fully articulated. More than that, it asked the right question—the one that had been lingering in the back of my mind: What is this sculpture trying to say?

This was the moment I realized that AI, while not a human critic, could serve as a powerful mirror—reflecting back insights that forced me to confront my artistic intent head-on.