Generative AI has historically been pretty bad at portraying Latin American culture accurately. Even though many of the resources and labor that power this technology come from this region, the resulting models keep replicating stereotypes and biases, failing at generating believable depictions of places like my home country of Chile.
This is what motivated
@tiagosdesign and me to develop "Artificial Us", an interactive installation and research project critically examining parallels between historical Latin American resource extraction and AI's neo-extractivism.
To do this, we fine-tuned multiple AI models with personal photos of our home countries, to then create 3D point-cloud environments that can be navigated in real time using
@touchdesigner . Body-tracking allows participants to explore these scenes, embarking on a narrative journey that exposes resource exploitation cycles.
"Artificial Us" is a collaborative thesis project by Matías Piña and Tiago Aragona, developed as part of their MFA thesis at
@parsons_dt . It questions if generative AI can capture the 'essence' of underrepresented cultures while exposing asymmetries in its development and deployment. The installation highlights how Global South resources fuel systems that often marginalize the very cultures from which they originate. It was exhibited at
@msftgarage , NYC, and at
@parsonsschoolofdesign .
Special thanks to
@samlavigne and
@ethan.silverman.7 for guiding us throughout this whole process!
