AI-generated combined-object sculpture

view full
paper here!

wavingflag--ar15-3e107{Object Americana}

I am an artist working at the intersection of Machine Intelligence, digital manufacturing, and 3D space. Object Americana is a project that seeks to define a new medium of sculpture. This genre is defined on the terms that exist between my self and the Machine agents I have created to partner with. I have spent over a year choreographing and refining the process by which we work together, with the ultimate goal to create pieces of physical material that are both fundamentally of the Machine and fundamentally of the Human.

The first sculptures of this new genre are presented here. Machine-Human collaboration has allowed for an uncanny new ability:
combining two or more objects into a new hybrid form, a type of feral alchemy with strange semiotic implications.

This project extends and modifies 3D Point Cloud Generative Adversarial Network Based on Tree Structured Graph Convolutions, Shu, Park, and Kwon, 2019.

BIG NEWS! Object Americana featured in a gallery exhibition!

Left Unsaid, Miss (in) Personae is a group exhibition created by Parsons School of Design's Fine Arts Exhibition Committee in partnership with COPE NYC. Press release here.

Left Unsaid is centered around the alienation from the physical world that we have all experienced throughout these endless months of global pandemic. '...words are at constant risk of misinterpretation in the cyber-network of the pandemic world. Missing in personae, words become relegated to other modes of communication such as suble gestures and visual cues.'

On view through July 18th, sign up here for a viewing slot! Located within the historic Pfizer building in Bed-Stuy, Brooklyn.



Object Americana currently comprises four sculptures created to represent new icons of America's past, present, future, and totality. Each of these sculptures was produced through a Machine-Human collaboration; the Machine combines the objects using a point-cloud GAN, and the Human takes the resulting point-cloud and produces a workable 3D object directly from it.

The antecedent objects to be hybridized were collected using a crowd-sourced process of polling data. A group consisting primarily of immigrants to the USA were asked about their thoughts on these three epochs of mental time, past, present, and future, and then about what objects they associate with those separate timeframes. As the polling data was analyzed, it became clear that there were a few more object-datapoints that spanned all three epochs; this realization resulted in the totality category, where the flag of the country and guns dominated. Thus, it can be said that the hybrid, combined-object flag--gun is the summation of American semiotic reality.

Each object has its own detailed page on this website; access them below. If you want to hear more about the theory and process behind this project, click here. For a more in-depth look at the polling data mentioned above, another webpage is in progress. All spinning 3D models on this site are interactive, in order to simulate the feeling of seeing a sculpture in person.

liberty--mcds-1e117{Our Lady Of Liberty And The Perpetual McDouble}


trump1--mcmansion-2e49{Excavation Of The Contemporary Nero In The Ruins That Birthed Him}


guillotine--ebike-2e125{Device Pertaining To The Situation We Find Ourselves In}


wavingflag--ar15-3e107{Object Americana}





I've had the pleasure of giving multiple presentations of Object Americana recently! The project has garnered attention from folks in the AI industry, the fine art world, and academia. Below is a selection of my talks on Object Americana, the issues it raises, and the "collaborative model" of working with AI I've developed to create the work.


The inaugural lecture of the project, this is the presentation I gave at my graduating class symposium, Parsons MFA Design + Technology HINDSIGHT Thesis Show 2020. Here I give a brief but detailed overview of the motivations and conceptual underpinnings of the project, touching on technical aspects as well. This lecture led off the second night of symposium, themed "Computation as Material."

You can also check out timestamp 1:46:06, to listen to a great Q&A session led by Dr. Pinar Yoldas, an architect, artist, and professor at University of California San Diego. She mentions the concept of "greebles" and asks my thoughts on how important it is the final objects be recognizable.

theCAT.io Creative Salon - Designing for AI & ML

I heard about Olivia Seow's designer-focused AI Creative Salon, and submitted a short talk on the methodology of Object Americana. It was approved, and I joined the prominent ranks of Drew Volpe and Abhay Agarwal, both of whom are making an impact on the industry side of the Machine Learning world. Volpe is a founder of First Star Ventures, and Agarwal leads Polytopal, a design firm working with cutting-edge AI strategies.

For this presentation, I was more focused on the "collaboration model" of working with Machine Intelligence that I developed throughout the course of creating Object Americana. This model serves to clarify the relationship between Human and Machine, and is based on openness and honesty about process. Event write-up here.