Good morning. I'm Auriea Harvey. I started making art on the internet in the nineties, then designed videogames, and for the last few years I've been making sculpture from software. I model a form digitally, and then it has to become an object. Deepening this practise and questions about how sculptures are fabricated today are my research project. Most of my work, I print on my own 3d printers and my sculpture involve a lot of personal handwork. But...

I've now had three major experiences of fabrication involving other hands. The first was my first marble piece, made by Monumental Labs in New York. While it was being made I never saw it in person or touched it, and when they asked me what I wanted, I didn't know how to answer as I didn't speak the language of stone. All decisions were made via Whats App photos and video calls. At the time, this was a deeply unsatsifactory experience for me. But it did fulfill a desire I had to have a sculpture materialize someplace I could not go. Something deeply interesting to me because I believe digital sculpture solves the often inconvenient demand of having to move my body from one country to another, or having to spend a lot of money in shipping, money which I would prefer goes into the making of the sculpture itself.

The second was Edmonia Triumphalis. The figure was made by Johnson Atelier, at factory scale, impersonal and standard, and the cost and weight meant I had to compromise on the material: urethane foam with a hard coat, not stone and bronze as I'd wanted. But the installation itself, in Bozeman, Montana, was different. The fountain basin and the Triumphal arch are made of rammed earth. Basically the material is the earth itself. Tinworks Art commissioned it, and with Montana State University students, a local architect and a producer on their team, a whole community built it. I wasn't there, and I still felt part of it. I did go see the sculpture in progress in New Jersey and for the final install and opening festivities in Bozeman, and I'm glad I did. The teamwork made the dream work. But it took 2 years, the producer's sanity, and funding from the Via Art Fund to make it happen.

The third was Mother/Child, milled by robot at the Digital Stone Project in Garfagnana, in Tuscany, and finished by my hand. I was there, and I touched it. Working the marble myself is where I learned how to more directly tie the digital file to the process of CNC milling and it's where the questions behind this research began. It's also where I wrote my first vibecoded apps, Tincture, Metope and Spolia.

Metope is a relief-carving studio that runs in a browser. Its brushes displace a height map, so what I paint is real depth that I can light, mill or print. Spolia is a collage plane, where scanned fragments sit in 3D space and weld, carve and overlap according to proceedural aesthetic rules, it is based on the way antique fragments or "spolia" were reset into twalls in Rome. I built both with Claude, and it fulfilled not only my desires to conjure up software at will, freeing my mind from decades of servitude to the machine, but also made me fascinated by what idosynchratic artist made software tools will mean for digital art going forward.

So my question has been this: what if software could carry knowledge between a fabricating artisan and me, so a form made in my studio can be made somewhere else? What does a human network of artisans look like, that gets me and potentially others away from commercial fabrication and materializing work worldwide and at human scale? How can AI made tool chains be in service to that connection?

How can sculptors today collaborate and fabricate differently? My proposal is to use AI as a translation layer between digital craft and handcraft, so a form made in one place can be made by many hands in another, as true collaborators and not clients and vendors. My guiding principle is that human judgment must direct the tool, never the other way round.

The research lives in a wiki, built around seven questions. Some pages are still in progress. I'll take you through four.

First, the file. For me the file is the artwork. It's the model I spend months shaping, and every object comes from it. The figure for Edmonia Triumphalis was a file before it was anything else. It's also the least valued part of the process for the art world, which doesn't understand what digital art is or could be: people want the marble or the bronze and treat the file as a receipt. But it is my stance that most art today is made by digital processes in the mix, and most sculpture involce 3d processes at some point. And that the 3d artist is terribly underrated MVP of sculpture in the 21st century. Meanwhile, prompt based 3D AI is moving into exactly that place, the making of 3D forms. So I have to watch it closely as it comes for MY craft. So my position is to use it as a tool, not as a replacement. And if possible unite with other craftspeople who see their art minimised because of the hypecycles of the digital.

Second, The hands. I started by saving makers on Instagram (because lots of artisans look for work by posting there, as do I): I collected 33 posts from 29 accounts, clustered in Italy, Egypt, Iran and South Asia, but I found I had seen almost nothing in Latin America, Africa or East Asia. So I went looking further, using AI to help me search in other languages and put them on a globe. (Drag the globe to turn it.) This helped me to understand the types of hands I felt affinity for. None of these people have been contacted yet. This is a map of who I'd like to collaborate with based on what they show.

Third, institutions. I don't yet know which ones, and I'm trying to start from what I can give them, not what I want from them.

This is the hardest part of the research. The question isn't whether an institution will host a work, it's what role it should play. I see three models. A host shows the work where it was made. A co-producer funds and arranges the making. A broker connects me to local fabricators and then steps back. Which would you trust, and what would make each one extractive? I'll put that to the roundtable tomorrow.

Last, what is best for humanity. My guide is the encyclical Magnifica Humanitas, Pope Leo XIV on AI and humanity. I can't summarize it in a minute. What I hold onto is that human conscience and judgment must remain the ones directing any technical process. In practice that means a maker should never have to become a technician for this to work.

Thank you. Tomorrow's roundtable is about the institutional question, and I'd love your thoughts there too.