MA.S[j]IEN

An Architecture for Doubt

Artist Statement

Artificial intelligence systems operate under the guise of neutral, seamless mathematical perfection, yet they are structurally shaped by the narrow, hetero-patriarchal, and extractive ideologies of the datasets that birth them. In their drive toward statistical optimization, mainstream machine learning models actively flatten the nuanced, relational, and embodied realities of human existence. MA.S[j]IEN: An Architecture for Doubt confronts this algorithmic reductionism through deliberate sabotage, feminist refusal, and the material friction of lived experience.

Working across a hybrid methodology that integrates software development with physical making, I intervene in the generative pipeline using custom node-based workflows in ComfyUI. By training custom LoRAs on my own disrupted paintings, gestural lines, and fragmented forms, and by defining anomalous latent space parameters, I force the model away from glossy, hyper-real tropes toward induced hallucinations and deliberate glitches. Drawing on Legacy Russell’s Glitch Feminism, these digital failures become sanctuaries of non-performance that fracture the rigid binaries and normative aesthetics rendering aging, non-Western, and feminine bodies invisible.

This tension unfolds spatially across the gallery as an architecture of doubt and hesitation. Suspended at the centre, three-meter-high translucent chiffon banners drift among white sculpted gestural lines and the maternal form of the Ouma figure, framed by opposing walls of digital prints and heavily textured mixed-media paintings. Beneath them, an electronic sculpture of exposed PC cooling fans, driven by an Arduino, breathes erratic kinetic life into the translucent chiffon images. As the viewer walks through the installation, sightlines continually collapse and refract: forcing the viewer to question what they are looking at - the print on the cloth, the mark on the wall, or the space in between. Here, visual ambiguity becomes an embodied critique of algorithmic certainty, amplified by the polyphonic soundscape Ek is hier, kan jy my sien, where layered female voices resist reduction into clean data.

Rooted in the philosophical extension Dubito, ergo cogito, ergo sum (I doubt, therefore I think, therefore I am) MA.S[j]IEN builds an architecture that holds uncertainty not as an error, but as an essential human faculty. By exposing the voids within our digital systems through touch, breath, glitch, and gesture, the work invites viewers to know the world not through data, but through presence: asserting that to doubt the machine is ultimately to preserve the human.

This body of work was presented as part of the Unisa CONVERGENCE 2026 Visual Multimedia Art Graduate Exhibition, 14 November 2026 – 15 January 2027 at the Unisa Art Gallery.