Decolonization Archive reimagines an Islamic manuscript in which a meticulously painted insect is accompanied by a description so vague it erases scientific identity. Neither butterfly nor moth, the creature lingers in the archive as an unnamed thing, suspended between taxonomy and myth.

This project reconstructs the insect through TouchDesigner, coding, and AI, producing a digital twin that exists in a liminal zone, between fact and folklore, manuscript and machine, past and present. The work opens a dialogue between archival opacity and algorithmic transparency. In one format, a live AI model streams scientific fragments about butterflies directly onto the folio, writing over its silence in real time. In another, the piece appears as a video, layering these outputs into a fixed record of reinterpretation.

Neither format offers correction or closure; both reveal a hybrid identity that resists stable classification.

By laying bare the AI’s process, its data traces, hesitations, and missteps, the project speaks to the core concerns of Debox: transparency, memory, and process. Here, AI is not a sealed black box but a collaborator whose interventions are partial, situated, and fallible. The misdescribed insect becomes a metaphor for colonial erasures and institutional misclassifications, while the generative layer of AI enacts possibilities of revival and resistance.

Ultimately, Decolonization Archive confronts how knowledge is written, forgotten, and revived. It proposes a decolonial strategy for digital heritage, one that foregrounds uncertainty, embraces incompleteness, and renders visible the hidden labor of both archive and algorithm. In doing so, it transforms an overlooked folio into a polyphonic site of critical transparency and reimagination.

In AI’s Eye, the gaze of the machine becomes the new lens of creation. A mechanical eye follows a floating disk, an orbiting gesture that recalls the dance between observer and observed. The work transforms the act of seeing into a dialogue: the human no longer only looks but is looked at.

Here, vision is not passive; it is an algorithmic awareness that tracks, interprets, and learns. The eye’s movement echoes the invisible presence of artificial intelligence, its capacity to follow, to remember, and to reconstruct meaning through observation. In this interplay, the boundaries between creator and creation dissolve. The eye does not merely witness; it participates in the act of making.

The digital gaze becomes both tool and collaborator, a mirror that reflects our own act of looking back at us.