The DXARTS SoftLab is a studio and an online platform whose mission is to examine the role of workmanship in artistic research, to redefine the use of crafting in the post-digital era, and to explore the body as an interface of control and resistance. It is part of the Department of Digital Arts and Experimental Media (DXARTS) at the University of Washington in Seattle.

AI-pocrypha by Laura Luna Castillo

AI-pocrypha by Laura Luna Castillo

Combining Natural Language Processing, GANs and other ideas and methods, I’m interested in exploring some kind of speculative objects that can become vessels for the unexpected, hallucinated narratives that blur the boundaries between fact and fiction.

I found inspiration on Italo Calvino’s Collection of Sand and Inivisble Citites, together with the objects known as ex-votos, specifically a fascinating article titled: Wandering Testimonies: The Diaspora of Mexican Ex-votos (and Their Stories) briefly discussing the ex-votos from Mexico. These objects have been transformed and passed on from different communities, from a religious practice reserved to the Spanish nobility, their use was slowly reclaimed and appropriated by the general population in Mexico. Ex-votos would often serve to drive a very personal narrative, often deviating or depicting dubious miracles (fig.1) attributed to different Saints, they would be hanged in chuches, worn within garments or as charms (fig.2).


Fig. 1 Ex-voto from Puebla - “Nurse Bartolo Garcia was very unhappy with his job because he did not like giving enemas, so he invoked San Sosimo [possibly Saint Chosmas] and since he was transferred he gives thanks with this small retablo. Puebla 1927”

Within this research, the idea of incorporating a tactile interface to drive a generative narrative is one of the core elements. I’m interested in researching if having textures and reliefs that engages other senses, could co-exist with a virtual reality, AI-driven narrative.

Experimenting with textual prompts I was able to get AI-generated “maps” (figs. 3 & 4), from these imaginary maps, a basic outline was co-authored in order to use machine embroidery and conductive thread (Fig.5). On top of the machine embroidered map, other imaginary details were added by hand as well as incorporating a simple circuit and its components(Fig. 6 & 7). 

Fig. 3 AI-generated imaginary map - Volcano detail

Fig. 4 AI-generated imaginary map - Puebla in 1774

Fig. 5 Imaginary map - Machine embroidery

This initial experiment combines Spanish-NLP text, first as inputs to generate images and then as outputs within an interactive interface. The components of the project are as follows:

  1. Spanish-NLP text inputs for VQGAN, asking it to generate maps of an imaginary city.

  2. Modifying those maps into vector images and embroidering some of the maps topographies/details with conductive thread, in order to produce capacitive sensors.

  3. From the map created within the embroidery, start asking the Spanish-NLL model to tell me stories about the streets and places of the map, thus placing more capacitive sensors with conductive thread.

  4. The resulting embroidered ex-voto will be interfaced with circuits to be able to trigger the NLP narratives by touching or “traversing” the map.

Fig. 6 Hand and machine embroidery combined with circuit components.

Fig. 7 Embroidered map, circuit in the back.

Electronic Pirouettes: Sound as Movement in Circuitry by Ioana Vreme Moser

Electronic Pirouettes: Sound as Movement in Circuitry by Ioana Vreme Moser

Stitching Code by Hannah Twigg-Smith

Stitching Code by Hannah Twigg-Smith