Abstract
Denial Sage is a sound art piece, which speculatively engages with history,
bias and conflict through the mediated perception of machine intelligence.
It creates this abstract fiction to examine non-human cognition, to create a
stranger with which to confront us with a flattened past and to let the vagar-
ies and specters of its perception haunt us. The audio material of Denial Sage
was created by a generative machine learning model trained on highly bi-
ased datasets comprised of .wav files from such sources as conflict report-
ing and emergency service scanner recordings, which were post-processed
and shaped into a multi-channel pseudo-generative installation. While the
sounds are often divorced from familiarity, glimpses of something eerily
known does come through, questioning what pasts form our future, when
even algorithms can hear it.
