AI Art and the Productive Potential of Opacity

Abstract

Opaqueness has long been mobilised within art as a means
of negotiating and expanding our perceptions of visibility and
invisibility, the knowable and the unknowable. This chapter
develops the existing counternarrative to transparent AI by
extending Asbjørn Skarvåg Grønstad’s (2020) work on art
and opacity. Grønstad illustrates how key concepts within
contemporary artistic practices – the political, the aesthetic, and
the cognitive – speak to the productive potential and critical
value of opaqueness. This chapter applies these ideas to artworks
incorporating a machine learning component, exploring opacity
as: (1) holding political value in reclaiming cultural rights through
techno-heritage, (2) an aesthetic frame to account for more-than-
human knowledge and (3) a ‘cognitive value’, or creative affordance
for imagination.