Against a pure white ground, a 16×9 grid holds 144 synthetic faces. Each independently nods, shakes its head, blinks, and opens its mouth, performing the same repertoire of verification gestures while refusing to fall into sync. The image moves from near-stillness into dense agitation, like a field of molecules brought to life, before arresting face by face. Human movement does not culminate in communication but in a machine-readable state.

​

The Verification Matrix is concerned less with whether recognition technology is accurate than with how recognition remakes those it recognizes. The faces have no names, memories, or social relations; they retain only the expressions the algorithm requires. Opening the mouth, blinking, and facing the camera—gestures once belonging to communication and confirmation—are reduced to verification parameters. What is truly unsettling is not an external observer, but the way people actively produce their own legibility in anticipation of being recognized, gradually learning to become “human” in terms a machine can understand.

​

As part of a practice-led inquiry, the work is not an illustration of research conclusions but a set of experimental conditions that can be viewed, experienced, and reflected upon. Synthetic faces detach the question from the privacy narrative of specific individuals: they resemble humans yet belong to no one, exposing how biometric systems manufacture standards in the name of a “universal humanity,” exclude deviation, and translate difference into risk. The 144 cells are at once portraits and specimens, equally arranged individuals and a dataset compressed by the same normative order.