Page 85 - AC-2-3

Arts & Communication Computer vision in tactical AI art

purposes as aloofness or cynicism, leading to indifference, rights. However, these processes are substantiated not
104
distrust, or resentment. The cultural value of tactical only by the AI industrialists and investors but also by most
AI art practices depends on their adequate positioning, users willing to trade their civic duties for the benefits of
representation, and preservation, whose range of problems surveillance-based consumerism, comfort, or security. The
has not been solved despite extensive efforts. 97 implicit public complacency reinforces the inequalities
embedded in and served by automation and drives society
The most elusive source of tactical AI art’s issues is
its sociotechnical milieu, which simultaneously provides toward a risky reliance on brittle data-based classification/
8,10
prediction systems.
critical affordances and threatens to relativize, debase,
neutralize, or instrumentalize their outcomes. Since the Tactical AI art’s main challenges lurk in its theoretical
mid-2010s, the AI industry has been sponsoring and foundations. As Martin Zeilinger noted, media art
105
promoting AI art as a powerful marketing and public activism has been largely influenced by a media theory canon
6
relations instrument, and the mainstream art market has from the left, with authors such as Rita Raley, Alexander
107
106
absorbed it in its trading portfolio. Thus, many decorative Galloway, McKenzie Wark, or Critical Engineering
108
non-tactical artworks serve as spectacles of corporate AI Working Group. This canon draws upon Michel de
109
power, while tactical ones frequently get recuperated, Certeau’s conceptual distinction between strategy and
98
99
occasionally become mere reflections of their targets, or tactics in cultural contexts. De Certeau defines strategy as a
lapse into the mystification of technocracy whereby a class goal-oriented set of practices focused on instrumentalizing
of tech-savvy artisans acts on behalf of “the (lay) people” by the structural affordances of their environmental,
articulating a vision of individual freedom realizable from social, or technological substrates. The strategy serves
within the power structures of the information society. administrative agendas by drawing on system-inherent
100
The corporate sector leverages AI art’s sociotechnical control architectures, often to contain divergent elements,
entanglement to systematically assimilate and often for instance, through algorithmic labor or consumer
exploit activist practices for refining the normalization management systems. Conversely, tactical practices
or circumventing problems they identified instead of articulate anti-authoritarian and oppositional responses to
101
correcting them. For instance, works that unveil the strategically advantageous positions of dominance through
vulnerabilities of face recognition algorithms can stir actions that are responsive, fluid, and embedded in the
up their technical improvements under the existing systems they challenge. However, theoretical notions about
application logic. For all these reasons, even when the strategic vs. tactical dichotomy and the suitability and
proactively intervening, tactical AI art opens questions meaningful impact of digital activist operations within the
beyond its apparent contributions. Should an activist systems and by means of strategic power seem to be too
102
action end up (directly or indirectly) being used by the AI vague, permissive, compromising, or implicitly compliant
industry to enhance its profitable instruments or remedy when facing the extent and sophistication of modern
its public image without necessarily improving its techno- AI’s techno-economic regimen and its broader realm of
110
ethical standards? Could tactical art disrupt the corporate neoliberal info-capitalism. Specifically, they imply the
AI regimen with lasting and desirable social consequences, question: what compromises, vulnerabilities, and other
and how effectively can it incite or enhance government trade-offs come with deciding or accepting to operate
policies for accountability and regulation of private AI critically within (not against) the infrastructures, protocols,
businesses with global influence? and ultimately the rules of that realm? Many tactical
artworks, even if praised by academia or the art scene, fail
Tactical AI art shares these political vulnerabilities to instigate noticeable long-term changes primarily because
with its audience. Modern AI’s social impact is marked by their makers could not reach the optimal ratio between
the uneven performance of applied AI systems: in some defiance, bravery, cunning, ambition, and survival instinct
contexts, they are beneficial; in others, superfluous, absurd, in answering that question. Under present circumstances,
abusive, or dangerous. The causes of this performative it seems that such conceptualization of art activism is often
103
unevenness are not purely technical. They relate to a shady more effective in unwillingly aiding the strategic power
facet of AI tech’s commercial success, which concerns to simultaneously exploit (recuperate) and marginalize
substantial investments in, yielding profits from, and critical artworks than in allowing them to disrupt that
normalizing the technologies for data capture, analysis, power. There is no denying that successful critical AI
and monetization. The vast asymmetry between societal artworks are effective in raising public awareness about AI
cost and private gain in the inflated rollout of these issues. They usually do it with more panache and charm
technologies aids the neoliberal political backlash that than investigative journalism or other non-artistic forms of
de facto seeks to bypass or erode civil, labor, and human activism, but they mostly remain within that “signal layer”

Volume 2 Issue 3 (2024) 12 doi: 10.36922/ac.2282