The Models

Overview

The Models (2025) is an interactive single-channel installation by artist duo dmstfctn exploring truth, wonder and the quirks of generative AI through the Italian improvised theatrical tradition of commedia dell’arte.

Project details

Started: May 2024
Completed: February 2025

Audiences can interact with this project through a mobile-accessible website, where they can:

Select AI tendencies they wish to see acted out
Trigger corresponding masks to appear on stage
Drop controversial props into the scene to heighten tension
Praise or boo each scene, which can lead to early endings

Masks respond in real time, with dialogue dynamically generated by a system of AI models.

Concept, design and development: Francesco Tacchini, Oliver Smith (dmstfctn)
Additional game development: Jenn Leung
Original soundtrack: Hero Image
Leonardo HPC advisor: Antonella Guidazzoli
HPC data scientist: Luca Mattei
Curation and production: Sineglossa
In collaboration with: CINECA Vision Lab, ART-ER
Under the patronage of: Regione Emilia-Romagna, Fondazione Bruno Kessler.
Funded by: European Union - European Digital Deal (as part of Next Generation EU programme)
Additional funding: Leonardo Supercomputing Centre (in-kind computing budget)

Researchers' installation

The installation runs continuously and consists of an endless series of theatrical scenes featuring a cast of commedia masks acting in response to audience input. Each scene is rendered in real-time using a video game engine, with dialogues and voices generated by AI. The masks embody both their traditional archetypal, exaggerated traits and behaviours often observed in generative AI: making things up, being antagonistic, or being overly servile.