The Italo-British duo dmstfctn presents in Madrid The Models, an audiovisual and interactive installation that fuses avant-garde and tradition by improvising representations of the Commedia dell’Arte through generative artificial intelligence and real-time audience participation.
Developed with the support of the Italian organization Sineglossa, the piece explores the improvised, absurd, and sometimes deceptive nature of artificial intelligences. On stage, AI models embody the traditional masks —Arlecchino, Balanzone, Brighella, Colombina, Pantaleone, and Pulcinella—, figures associated with servitude, deceit, or confrontation, who interact and dialogue in real time, thus generating a unique theatrical experience.
With this project, SOLO CSV opens to the public for the first time its cinema, a new space located in the basement level, conceived for projections and audiovisual or interactive works, thereby expanding the scope of its artistic program.
Each mask in The Models is inspired by paper silhouettes created by Bolognese theatre companies in the early 20th century, preserved at the Ferrari Museum in Parma. The scenes revolve around symbolic props linked to superstition, folklore, or deceit, and unfold across 64 theatre backdrops painted at the beginning of the 20th century and sourced from the Museo dei Burattini in Bologna, depicting nocturnal, festive, urban, natural, and interior landscapes.
During an artistic residency with Sineglossa, which accompanied the entire creative process, the artists developed the dialogue-generation system based on LLM-type artificial intelligence models, using the Leonardo supercomputer. The residency and production of the work were supported, through this Italian institution, by the European Union’s Next Generation Funds.
THE ROLE OF THE AUDIENCE
The audience plays a key role by selecting, from their mobile phones, the masks and props for each scene. Based on these choices, the AI randomly selects a backdrop and generates the dialogues, allowing for up to 26,800 unique combinations. The result is a series of unpredictable performances, filled with conspiracies, mistakes, and surprises.
During the sessions, attendees can throw coins, flowers, tomatoes, or eggs as a form of feedback, evoking the participatory tradition of the Commedia dell’Arte.
ABOUT DMSTFCTN
dmstfctn (Oliver Smith and Francesco Tacchini) explore complex systems through simulation, fiction, and experience. They often employ game engines in the creation of their works as tools for shared learning, exploration, and critique.
Their research on AI folklore seeks to frame the anomalies and quirks of AI systems not as errors to be corrected, but as creative phenomena worthy of attention, fostering a more nuanced understanding of artificial intelligence.
Since 2018, dmstfctn have performed and exhibited internationally at institutions such as Serpentine, Berghain, HKW, Rhizome, Unsound, CTM, and transmediale. Their audiovisual work has been published by Mille Plateaux and Krisis Publishing, and featured in Ø by Hyper-dub (Flatlines). Their essay Large Lore Models, written with Eva Jäger and Alasdair Milne, has been published in the UK, the Netherlands, and China.
They have collaborated with leading artistic and scientific institutions, including the Leonardo supercomputer and the Alan Turing Institute. In 2024, they were nominated for the STAR+T+S Prize for their videogame Godmode Epochs, and won the Edigma Semibreve 2023 Award for their work GOD MODE (ep. 1). They are currently n-space fellows at Somerset House Studios in London.