Journal of Aesthetics and Phenomenology

ISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/rfap20

Future Imaginings in Art and Artificial Intelligence

Andreia Machado Oliveira

To cite this article: Andreia Machado Oliveira (2022) Future Imaginings in Art and
Artificial Intelligence, Journal of Aesthetics and Phenomenology, 9:2, 209-225, DOI:
10.1080/20539320.2022.2150467

To link to this article: https://doi.org/10.1080/20539320.2022.2150467

Published online: 21 Jun 2023.

Submit your article to this journal

Article views: 400

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=rfap20
JOURNAL OF AESTHETICS AND PHENOMENOLOGY
2022, VOL. 9, NO. 2, 209–225
https://doi.org/10.1080/20539320.2022.2150467

Future Imaginings in Art and Artificial Intelligence

Andreia Machado Oliveira
Federal University of Santa Maria/UFSM; National Council for Scientific and Technological Development of
Brazil/CNPq

ABSTRACT KEYWORDS
In considering future imaginings of art and Artificial Intelligence art; artificial intelligence;
(AI), we point out two key premises to our approach: the human- aesthetics; imagination;
machine relation without opposition between the two terms; and behaviours; practice-led art
practices
the implications of artificial intelligence within practice-led art
practices. According to Gilbert Simondon, there is a biological and
technological evolution that does not separate nature and technics
or culture and technics, since what is defined by human nature is
already part of a technological system. Yuk Hui criticizes the idea of
technological singularity supported by strong artificial intelligence
which converges on itself. However, there is also the possibility of
multiple bifurcations emerging from technodiverse practices and
aesthetics. On algorithmic art, Lev Manovich points out that AI is
increasingly culturally influencing our imagination, choices, and
behaviours and that it has built its own imagery and cognitive
patterns in society. Joanna Zylinska connects the dots between AI,
creativity and innovation, which by bringing AI to research into
practice-led art practices, opens the problematization of AI to the
possibilities of reprogramming the human sensory-cognitive
apparatus. In this paper, we extend these approaches to examine
whether these computer systems that enable the processing of big
data can or cannot program our imagination, desires, likes, sensa­
tions, perceptions and attitudes. Finally, in order to think the worlds
built through these emerging AI technologies, we ask what kind of
society we want to live in, and what roles we will allow ubiquitous
computational technologies, such as AI, to play in our daily lives.

1234
Coming from the current context permeated by technological systems of Artificial
Intelligence (AI), we discern a growing and recurring concern for the future, for its
images and imaginaries, whether foreseeable or unpredictable. From initial primitive
investigations in the 1950s to the present studies in neural algorithms, machine learning
(deep learning), big data, the internet of things, algorithmic art, metabolic art, and
generative adversarial networks (GANs), algorithms and data structures have become
more functionally complex in networks, interfering with cultural patterns and human
decisions, conditioning and anticipating behaviours, and affecting social, political, cul­
tural, and economic organization.

CONTACT Andreia Machado Oliveira [email protected]

http://lattes.cnpq.br/7243757837987821
© 2022 Informa UK Limited, trading as Taylor & Francis Group
210 A. MACHADO OLIVEIRA

By relating art and AI, artists explore the aesthetic possibilities within the types of data
mined in different contexts, inquire into what patterns can be expected, generated and
implemented through complex automation and problem-solving technologies, and,
mainly, problematize and seek to subvert the functioning of technologies. They provoke
us to think about how these technologies require a conceptual redefinition of what we
understand by concepts such as ‘human’, ‘species’, ‘machines’, ‘relationship’, ‘behaviour’,
and ‘emotion’. As Mulder writes: “we should study images and technologies not to find
out what information they might contain (knowledge) but what kind of capacity for
action they enable (agency).”5 Images and technologies are not only modified by the
exchange of information, but are also actors that produce informational patterns based
on what and how we think and feel, and by how we are induced to desire and to spurn.6
Through a practice-led art practice perspective, this paper thinks about how AI Art
practices can contribute to rethinking technological practices in artificial intelligence,
and vice versa. Media art theorists such as Lev Manovich and Joanna Zylinska, among
others, have problematized AI in the field of the arts. Manovich points out that AI is
increasingly influencing our cultural choices, behaviours and imagination and that it has
built its own imagery and cognitive patterns. Zylinska connects the dots between AI,
creativity and innovation, which, by bringing AI to research into practice-led art prac­
tices, opens the problematization of AI to the possibilities of reprogramming the human
sensory-cognitive apparatus.
As such, understanding that the future is reflected in the present and that we can
define it, we can ask ourselves what futures can or cannot emerge from present techno­
logical scenarios? We ask whether computer systems that enable the processing of big
data and deep learning so as to predict behaviours can or cannot program our imagina­
tion, desires, likes, sensations, sensitivity, perceptions and attitudes. Such questions incite
us to rethink our lived experiences and aesthetic production with AI, examining the
threats, possibilities and impossibilities that AI presents in a networked world.
Avoiding salvationist or apocalyptic positions in relation to AI, the paper sets out from
the premise that we need to review positions that put the human and the machine in
opposition, since behind this type of vision there is a slavocratic posture in which the
human and machine alternate between subordinate or dominant roles vis-a-vis one
another. Proceeding in this way, we invoke some authors who refer to assemblages
between humans and machines and reaffirm the incoherence of the opposition and
polarization of the human and the machine, such as Gilbert Simondon, Hui, Haraway
and Braidotti.
According to Simondon, there is a biological and technological evolution that does not
separate nature and technics or culture and technics, since what is defined by human
nature is already part of a technological system.7 Hui criticizes the idea of technological
singularity and the opposition of machine and human so that the machine dominates and
replaces the human. Haraway and Braidotti have prompted us to rethink what we
understand as human in terms of assemblages with machines and other species.8
Thus, in this paper, we point out two key premises to think about art and AI: the
possibility and desirability of conceiving of the human-machine relation without opposi­
tion between the two terms; and the implications of AI within practice-led art practices.
We position the importance of art as conjunctive with several areas of contemporary
knowledge and examine how, through its projects, it problematizes AI by keeping open
 JOURNAL OF AESTHETICS AND PHENOMENOLOGY 211

imaginable and not-yet imaginable futures of diverse and plural interconnected worlds.
From this perspective and through the integration of AI, the key point of this paper is to
show how art practices can allow us to reimagine both the extension and the intension of
our concepts of ‘art’ and ‘AI’. On the one hand, each technology that appears brings with
it the need to review what we understand by art – as occurred with the technologies of
image reproduction9; on the other hand, art opens to the possibility of multiple bifurca­
tions emerging from technodiverse practices and aesthetics, by demonstrating in its
practices the intrinsic assemblages between technology, aesthetics and concepts.10
When thinking about art and AI, we observe that the constructions of imaginaries and
inventions emanate from the assemblages between humans and non-humans with their
associated milieus.

Human-machine agency AI
There is no human-machine opposition for Simondon. According to him, technology is
not posed in terms of the tool/instrument, but as a gesture inserted within technical
culture. In some way or other, we have always been immersed in technical culture. We
subject ourselves to technologies, we produce them and are produced by them — what is
defined as human nature is already part of a technological system. Simondon seeks to
integrate technology and culture, pointing out that it is necessary for the technical object
to be known and familiar so that the relationship between man and machine is stable and
valid: hence, the need for a cultural understanding of technology and technics. He
criticizes approaches that understand technics in a purely utilitarian way, devoid of
sense and meaning, since knowledge of human culture necessarily involves knowledge
of technologies. With reference to Simondon, Hui states that technical reality

inscribes in itself also the human reality, not only because technology is the realization of the
mental schemes conceived by the designers and engineers, influenced by social and political
structures of the human society, but also because the technical reality equally transforms the
human reality.11

From another perspective, Haraway emphasizes our composite multiplicity not only by
our being identified as human but through the diversity of bodies we create as non-
human agents, including machines and technologies. Haraway addresses the interaction
of the artificial with the natural; for her, artificial life, transgenic art and robotics
problematize the body, life and human being.12 Haraway tells us about non-human
agents or contexts, in which machines and their technologies are included, where
technologies:

are composed of diverse agents of interpretation, agents of recording, and agents for
directing and multiplying relational action. These agents can be human beings or parts of
human beings, other organisms in part or whole, machines of many kinds, or other sorts of
entrained compounds.13

By emphasizing that we are a composite species of ‘permanent becoming’, transiting

from the Anthropocene to the Chthulucene, Haraway helps us to realize that what we call
human is a composite of humans and non-humans, a mixture of viruses and bacteria, of
212 A. MACHADO OLIVEIRA

plants, minerals, machines and software – the latter increasingly incorporated into our
system.14
Corroborating Haraway’s ideas, critical posthumanist Braidotti also problematizes the
human through non-human becomings. Pitting the notion of a dominant subject (male,
white, European, urban) and an anthropocentric view, the author places post-humanism
as a convergence of anti-humanism and anti-anthropocentrism, in a way that goes
farther in a more complex and progressive direction. Braidotti identifies an ethical aspect
to Haraway’s speculative undertaking: “The ethical part of the project concerns the
creation of a new kinship system: a new social nexus and new forms of social connection
with these techno-others.”15 She points to the need for a post-humanist approach that
does not divide the given of nature and the constructed of culture. On the contrary, she
addresses the contingent historical aspects in the relationship between nature and
culture, considering vital, self-organization and non-naturalistic structures.16 Braidotti
underscores the importance of rethinking subjectivity as a collective agency that encom­
passes human and non-human actors, as multiple compositions and collective practices
of a post-humanist-becoming.
The post-human is conceived by Hui from the idea of an inorganic organicity. In his
book Recursivity and Contingency, Hui historically investigates algorithmic contingency
and brings an elucidation of the contemporary situation in which we live, reviewing the
inadequate polarization between nature and technics. For Hui:

The cybernetic machines, especially the Turing machine, have a new status, since it is no
longer a mere mechanism in the Cartesian sense, nor is it a living being. Instead, it is an
organo-mechanical being: a mechanical being implemented in an organic form. The organic
form doesn’t just give form as such, but also organizes matter conditioned by a definite end
or a relatively open end inscribed in the recursive algorithm.17

From a historical-critical approach to the theorization of the organic in current

philosophy, Hui points out two major lines of thought in the 20th century: organicism
(ecology and cybernetics) and organology. Organology deviates from anthropocentric
and substantialist approaches to the human and to culture by introducing the acci­
dental and the contingent, and is more concerned with posthuman than transhuman
conceptions. By stating that the human is a technical existence, Hui underscores the
importance of studying the man-machine relationship from a non-anthropocentric
outlook.
Thus, from a different position to transhumanist ideology, in order to think the
implications between organic and digital machines, Hui has insisted on the importance
and urgency of bringing technical existence to the technological, political and social
debate, in order to cast a new look on the hegemonic technocratic discourse of techno­
logical imperialism. Hui writes about post-humanism in order to avoid a homogeneous
technological development based on a Eurocentric, deterministic, apocalyptic and tota­
litarian vision. He highlights the diversity of understandings of post-humanism and of
the human: “The concept of the human is a contingent historical concept. We are already
posthuman when we subscribe to the view that the human is a technical existence.”18 He
advances a post-humanism that does not dissociate the organic and inorganic. In
contrast, he points out that transhumanism takes refuge in technology as the savior of
 JOURNAL OF AESTHETICS AND PHENOMENOLOGY 213

the human, with a “desire of inorganic immortality.”19 There is a deterministic emphasis

on automation systems in our society, which as Hui warns:
[. . .] will reflect on the increasing determination of technical systems realized in the new
wave of industrialization, fueled by artificial intelligence, machine learning, and all sorts of
surveillance technologies endowed with a transhumanist ideology that wants to overcome
the limit of the human and politics.20

Hegemonic discourses seek to subjugate local cultures in order to promote the strength­
ening of universalizing technologies. The current policies which endorse automation
uphold a capitalist and colonialist policy of automatism that sees competitiveness as
a solution. Hui criticizes the idea of right-wing accelerationism (as a continuation of
Enlightenment thought), such as that espoused by Nick Land, who proclaims the
singularity of technological development in favor of the intensification of capitalism in
the perspective of its self-transcendence. Technological singularity, supported by strong
artificial intelligence, prophesies that we will reach the level of development of an
omnipotent and omnipresent super-intelligence, superior to all others, which converges
on itself. And because it is so specific and separate from the human, it becomes unique
and transcends this category. According to the right-wing accelerationist view, only this
hegemonic super-intelligence would be able to manage the manifold economic, political
and social conflicts of humanity on a global time axis, since they would be synchronous
and therefore predictable.
We must instead escape its global time-axis, escape a (trans)humanism that subordinates
other beings to the terms of its own destiny, and propose a new agenda and imagination of
technology that open up new forms of social, political, and aesthetic life and new relations
with nonhumans, the earth, and the cosmos.21

Within the networked global synchrony of modern technologies, there is a convergence

of time causing an explosion of intelligence, but also of human servility. However, there is
also the possibility of multiple bifurcations arising from this convergence;
[...] bifurcations emerging from localities and their cosmotechnics, from technodiverse
practices, such as the technological practices that have been developed in the field of art.22

For Hui, transhumanism aims to explore technology at its limits, prescribing it as

a means of saving the human, which it considers obsolete:
[Transhumanists] embrace functionalism (seeing the human as composed of functions that
can be improved individually) and an interdisciplinary program for human enhancement,
including information technology, computer science, cognitive science and neuroscience,
neural- computer interface research, materials science, artificial intelligence, regenerative
medicine and life extension, genetic engineering, and nanotechnology.23

As such, when talking about the interconnections between post-humanism and artificial
intelligence, we point to the pertinence of interdisciplinary research on AI, inclusive of
art, philosophy, science and technology, seeing how we are being “moulded” and
modulated constantly through discursive, behavioral and aesthetic patterns arising
from AI practices. When referring to AI, Hui suggests that we need to consider the
diversity of approaches to what we understand by intelligence, requiring a historicity of
the term intelligence and of artificial intelligence itself. For Hui machine intelligence is
214 A. MACHADO OLIVEIRA

more susceptible to mutations than human intelligence, since it occurs in the indetermi­
nacy of machinic and human agency, within recursive digital operations.24 On his
account:

. . . to expose the limit of artificial intelligence is not to render machines weak again, but
rather to liberate machine intelligence from the bias of certain notions of intelligence, and
therefore, to conceive new political ecologies and political economies of machine
intelligence.25

In the case of AI, intelligence is capable of producing “transubstantiations”, when

transferring from one material to another, through mutations that are broader and faster,
and, in this sense, practically unlimited.26
Problematizing artificial intelligence more pointedly, Meredith Broussard, in her book
Artificial Unintelligence: How Computers Misunderstand the World (2018), cautions us
against a certain blind optimism in relation to AI by taking a critical stance on the current
production and uses of AI. She differentiates between programs that are “popular” and
those that are “good” or beneficial, to warn us that those that are widely used are not
necessarily always well-intentioned. In this sense, the author points to the need and
urgency of reviewing computational processes and decisions, since, given the current
potential of AI, we have a responsibility to ask how computers interpret data, and when
data is misinterpreted, to determine what is behind the erroneous programming. With
large-scale automation, made possible by the combination of artificial intelligence and
digital networks, we can discern algorithms of oppression.
In order to think about the worlds we have built with emerging technologies, several
theorists, such as those mentioned above, bolster our concerns as to what kind of society
we want to live in, and the direction we are giving to digital technologies, such as AI, in
our daily lives, as well as in systems of governance and control of bodies and subjectiv­
ities. In this sense, AI, as an emerging technology, is causing us to review how we feel and
perceive the world, and reconsider what aesthetics are being built in a world with AI. As
Walter Benjamin posits, the technological means of art production are not merely devices
foreign to creation, but determinants of its aesthetic procedures and processes.27 Thus,
we question how AI has changed current aesthetic standards and our ways of perceiving
the world, and probe into how art, by appropriating AI, has problematized AI itself and
provoked us towards other technological incorporations and, consequently, explore
other aesthetics and their transgressive capabilities.

AI Art and aesthetics

It is important to note, building on what was said in the previous part, that technologies
that use AI invite us to redefine the very notion of the human as a multiplicity of
heterogeneous modalities capable of recursive reciprocities where all involved partici­
pants, human and non-human, are not only modified by the exchange of information but
through the manner they produce informational patterns as data. In this sense, media art
theorists have problematized AI within artistic practices, inquiring into how forms of
machine learning, for instance, have affected us by altering our sensory experience and
aesthetic sensibility.
 JOURNAL OF AESTHETICS AND PHENOMENOLOGY 215

By altering our sensory apparatus and perceptions, AI productions raise new aesthetic
and conceptual questions that are brought to artistic practices. From a practice-led
research perspective, we come to think about the implications of artificial intelligence
within practice-led art practice, which includes how practices in the arts have guided,
influenced and conducted new research and productions in AI. By problematizing,
researching and producing within AI, artistic practices have guided and contributed to
the understanding of AI practices in various fields of knowledge. For Candy, practice-led
research occurs when “research leads primarily to new understandings about practice.”28
Practice-led research addresses the nature of practices and their operational meanings,
aiming to advance knowledge about/within practice itself.
When considering how practices in AI Art have contributed to the knowledge and
problematization of AI itself, we bring in the positions of media art theorists Lev
Manovich and Joanna Zylinska. On algorithmic art, Manovich refers to a cultural AI
that is available on devices and services that reach billions of users, identifying four
content categories: content selection, content segmentation, assistance in creating/edit­
ing new content, and fully autonomous content creation.29 Going beyond the automation
of cognition and serving as a medium for artistic imagination, AI has built its own
current imagery and cognitive patterns. According to Manovich:

We need new methods for seeing culture at its new scale, velocity, and connectivity that can
combine both qualitative and quantitative approaches and that can reveal the full variability
of this new ecosystem without reducing it to a small number of categories.30

Manovich warns us of a gradual automation of aesthetic decision-making, questioning us

about the reduction, or not, of aesthetic diversity today. Regarding diversity, referring to
his Selfiecity project, he emphasizes that expressions and interactions are unique; even
though there are many facial patterns, there are unique formal and gestural character­
istics, and it is interesting to analyze such differences. He raises some pertinent questions
regarding art and AI: “Does AI integration in cultural production and reception lead to
a decrease in aesthetic variability? Or does it, on the contrary, increase it? And what are
the different ways in which aesthetic variability can be defined and measured?”31
Manovich understands that this issue of aesthetic variability and classification parameters
can lead us to think about other approaches to aesthetics in the digital age and possible
future reduction of diversity in cultural expressions created by automation technologies.
In the book AI ART, Machine Visions and Warped Dreams, Zylinska connects the dots
between AI, creativity and innovation. In bringing AI into art, Zylinska contextualizes art
from a post-art-historical position of media theory, considering conceptual, historical,
social and technological practices. When conceiving art in a relational way, she argues
that a philosophical approach beyond aesthetics is essential, and requires the considera­
tion of issues of labour, robotisation and the long-term survival of the human species
(Zylinska, 2020).
In the same direction as Simondon and Hui, Zylinska holds that art has always been
linked to technics, and states that it can, to this degree, already be considered artificially
intelligent: “I propose to see different forms of human activity, including art, as having
always been technical, and thus also, to some extent, artificially intelligent.”32 She asks:
“Is there an ontological difference between early computer-generated art, net art and the
more recent forms of AI-driven art?”33 Zylinska presents a critical overview of the
216 A. MACHADO OLIVEIRA

concepts of intelligence and AI, raises ethical issues pertinent to AI, and considers “the
sociopolitical and psychopolitical risks of redesigning the artistic apparatus.”34 From her
critical viewpoint, she proposes a relationship between AI and ‘Anthropocene
Imperative’ AI, referring to the planetary climate crisis. By bringing AI to research into
artistic practices, it opens the problematization of AI to the possibilities of reprogram­
ming the human sensory-cognitive apparatus, since art works directly with the senses
and sensibilities. Because art and technology have always been involved with one another,
for Zylinska, there is no polarization between them, and it is necessary to look at how
they reciprocally constitute each other in different practices and cultures, towards a better
AI discourse. “The main premise of this discourse, I suggest, would not pitch the human
against the machine but would rather adopt the human-with-the-machine, or even, more
radically, the human-as-a-machine scenario.”35 From Simondon’s perspective of think­
ing of technologies as social gestures, Zylinska also considers AI as a social agent that
regulates tastes, the market and defining visuality.

Gradually a new sense of ‘being human’ is therefore emerging, consisting of gestures, bodily
movements, voice and language affectations, needs, desires and preferences drawn from the
multiple data available online and then transmuted by the deep learning networks into what
counts as ‘the human experience’.36

In addition to media art theorists, artists have explored AI in their productions in various
ways and changed the critical conception of what artistic creation can be. Blaise Agüera
y Arcas, in the article “Art in the Age of Machine Intelligence”, states that technologies
have always affected art, such as the invention of applied pigments, the written press,
photography, computers since their invention, and, in an intensive way, AI today. For
Agüera y Arcas, AI can expand artistic possibilities, through the expressivity implicit in
its technology.37
Within the diversity of current production in art and AI, we can mention other artists
as active in AI Art.38 London-based Turkish artist, engineer and computer scientist
Memo Akten stands out as an artist, researcher and philomath and the scope of his
work bears closer examination as a demonstration of the interaction of AI and art.39
Akten has developed AI Art through a variety of artistic proposals, such as experimental
films, large-scale responsive installations, and performances. In his computational art­
work You are what we see, Akten works with inputs that are processed through AI filters
to become transformed outputs, in order to make the point that we always see through
filters. In We are all made of stardust, he uses images of the cosmos as a reference for the
generation of datasets in his artwork, and breaks away from the usual sourcing of data
based on human activity. In his Pattern Recognition, two human dancers and eight
robotic spotlights dance together, with the system sensing the dancers’ movements,
controlling the movement of the lights and choreographing the two pieces of information
into patterns generated by the interaction of the dancers’ movements and the robotic
spotlights. In this proposal, creation takes place in the integration of artist and machine,
as a human-machine assemblage, where the machine has autonomy as an element of
creation in the performance despite functioning as an extension of the artist.
Based on the role of AI in its proposals, Akten states that machines always produce
behaviours, but not necessarily art; they iteratively fashion solutions to problems, yet
there is always human intent. With his artistic proposals, he questions what it means to
 JOURNAL OF AESTHETICS AND PHENOMENOLOGY 217

Figure 1. Learning to See: Gloomy Sunday/Memo Akten (TR) by Ars Electronica “Learning to See:
Gloomy Sunday/Memo Akten (TR)” by Ars Electronica is licensed under CC BY-NC-ND 2.0.

learn, understand and remember and creates within the challenge of designing systems
with intentionality. Inspired by the human visual cortex, Learning to See (Figure 1)
consists of a series of works comprising video and interactive installation using machine-
learning algorithms as neural networks. In this project, AI technology — which is no
longer modeled after the human mind – finds itself reflected back onto profoundly
human ways of perceiving and allows us to explore the “difficulty of seeing the world
through the eyes of others.”40
In the video Optimising for Beauty, Akten works with portraiture and shows an
artificial neural network in the process of being trained on a dataset made up of
thousands of portraits of celebrities. The artificial neural network generates fictitious
faces and learns to create a form of idealized beauty in which individual features are
gradually dissolved into one single face that averages differences. Akten intentionally
searches for this unsettling uniformity with an MLE (maximum likelihood estimation)
Deep Learning algorithm. “The result is a disturbing set of faces who all seem to be the
same person, none of whom actually exists.”41
In contrast, the American artist Trevor Paglen thinks in a distinct and critical way
about how computers “see” through devices such as powerful telescopes, and questions
how machines see, monitor and analyze images to understand the mechanics of this
seeing and its social and political implications. Understanding that vision is always
a cultural, historical and political construction, Paglen points out that there are “invisible
images” (Figure 2) that humans do not have access to: images that belong to databases
with an unfathomable number of images for humans; images created internally by
computers themselves in order to generate a solution; or the unseen automatic images
of our subconscious captured in our daily lives without our realizing it. Paglen’s propo­
sals provoke us to ask who benefits from these technologies and for what purposes they
serve within military and political systems, corporate enterprises, marketing and adver­
tising concerns etc.
218 A. MACHADO OLIVEIRA

Figure 2. Machine Vision facial recognition detail, art by Trevor Paglen “Machine Vision facial
recognition detail, art by Trevor Paglen, Sites Unseen, San Diego Museum of Contemporary Art, San
Diego, California, USA” by gruntzooki is licensed under CC BY-SA 2.0.

In another proposal, the exposition From ‘Apple’ to ‘Anomaly’ (2019-2020), Paglen

printed out some 30,000 photographs from ImageNet to show how computers see and
recognize images using training sets, thereby raising “questions about the inhumane
aspects of some of the viewing practices installed by the current political regimes yet
hidden from general view”.42 The artist aims to explain which are the parameters for
the construction of patterns of images recognized by computers, as well as how such
image patterns influence social decision-making. Paglen’s work “does not amount to
producing political art per se, but it does engage our sense and sensibility to reprogram
the human cognitive-sensory apparatus – and maybe to open it up to a different
hack.”43 In yet another proposal, Paglen led ImageNet’s project Person (2019), which
has an app to upload images to show how images of people are classified based on
ImageNet’s more than 2,800 person categories and to demonstrate how this effort is not
unbiased.
In fact, there are no neutral categories in ImageNet, because the selection of images always
interacts with the meaning of words. The politics are baked into the classificatory logic, even
when the words aren’t offensive. ImageNet is a lesson, in this sense, of what happens when
people are categorized like objects. But this practice has only become more common in
recent years, often inside the tech companies.44

To elucidate the biases behind the programming of AI by developers, American artist

Lauren Lee McCarthy developed the performance LAUREN in which she becomes
a version of Amazon’s virtual assistant Alexa to show how AI guides decisions and steers
personal choice.
The artist aspired to be better than an AI, because she could understand them as a person
and anticipate their needs. With this project the artist reveals how we are willing to exchange
intimacy for convenience and allow a set of algorithms to control our lives.45
 JOURNAL OF AESTHETICS AND PHENOMENOLOGY 219

Figure 3. Conversacube/Lauren Lee McCarthy (US) by Ars Electronica “Conversacube/Lauren McCarthy

(US)” by Ars Electronica is licensed under CC BY-NC-ND 2.0.

McCarthy sets up shop in participants’ homes for 24 hours to assist them any way she can
to one-up Alexa’s smart home intelligence. “The artist explores human-machine inter­
action through the lens of playfulness, irony and pastiche, demystifying the grand
promises behind some of AI, while also raising the question of the willing suspension
of privacy – which only starts feeling weird when the spying is undertaken by a ‘human
Alexa’.”46
In Conversacube (2010), another work by McCarthy, a metal cube that serves as an
interactive conversation steering device, produces politically correct conversations
between a group of people. Moderated and guided by AI, the cube listens and directs
the conversation based on keywords (such as accept, agree, smile . . .) so that
a comfortable and smooth conversation is established in which there is no room for
conflict, mistakes, anxieties, or fears. McCarthy offers the Conversacube (Figure 3) as an
equivocal proposal, in that on the one hand it can be seen as

an actual commercial product that uses technology to improve interactions, and on the
other hand, is critical of our dependence on technology and choreographed social routines,
hinting at a dystopic future where we sacrifice our autonomy to avoid having to face
anything uncomfortable.47

The device is an ambiguous object in that users interact with the device in public
spaces, such as cafes and shopping malls, unsure whether the cube is a new commercial
device or an artistic gag.

AI, Art and Future Imaginings

With the purpose of exploring and problematizing AI, art invites us to review its own
aesthetic parameters. Techno-aesthetic objects are an extension of the natural and
cultural world, which as a point of convergence and their positioning will be through
an aesthetic gesture. Aesthetics makes it possible to go beyond primary technological
procedures, just as technology is the means by which a certain aesthetic is produced.
220 A. MACHADO OLIVEIRA

Aesthetics and technology are linked by a continuous spectrum, while Art is constituted
in the midst of technology and aesthetics, knowledge and sensation and perception.
Like Manovich and Zylinska, many artists have concerns about AI in art with regards
to its social, political and ecological aspects. Despite AI still only being capable of
providing a very basic level of intelligence focused on information manipulation and
pattern recognition, on a social and political level, we need to understand the increasing
agency and functionality of human/machine assemblages under a new light. For Luciana
Parisi:

The cybernetic network of communication has not only absorbed physical and cognitive
labour into its circuits of reproduction, but is, more importantly, learning from human
culture, through the data analysis of behaviours, the contextual use of content and the
sourcing of knowledge.48

Through the reflections of the philosophers and artists mentioned above, we need to
consider future imaginings of our existence as inescapably conditioned and modulated
by AI. Engaging such future imaginings means coming to terms with a networked system
that occurs from the associated agency between the human and the non-human as
heterogeneous individuations from which novel social behaviour patterns result as new
modes of social understanding and experiencing of the world. To this end, we need to
complexify and nuance our ethical, epistemic, and ontological positions to reconsider the
human/machine relation, all too often exclusively articulated through reductive dis­
courses of enslavement or alienation49 – sometimes of the human, sometimes of the
machine – and which diminish worlds and existences to the calculable and predictable.50
Hui comments that AI constantly evaluates the past to anticipate the future, which in
turn determines the present. “Human beings are reintegrated into the temporality of
machines, not only as individuals but also as collectives and communities. This is
precisely what is called algorithmic governmentality.”51 In a critical sense, Hui stresses
the importance of a vision of technodiversity that includes and links technologies with
cultural diversity. Apprehensive about the ethics at play, he writes that “The challenge of
artificial intelligence is not about building a super-intelligence, but rather a matter of
facilitating noodiversity. And for noodiversity to be possible, we will need to develop
technodiversity.”52 As such, the ethics of technology aspires to different modes of knowl­
edge – such as the posthumanism – that open up technodiversity and future technolo­
gical development, surpassing the dominant normativities that govern certain subjects of
knowledge.
In agreement with Hui, Braidotti criticizes transhumanist positions that bet on the
technological imperative as an absolute solution to humanity’s problems within an
apocalyptic and unidirectional vision. She sees posthumanism from
a multidimensional perspective that does not separate nature, culture and technology,
conceptually posited as a nature-culture-techno continuum in which zoe-bios-technology
converges. Considering future possibilities, she points to the need for a “nomadic sub­
jectivity”, for “active transposition, a transformation at the in-depth level, a change of
culture akin to genetic mutations, but registered also at the ethical level. In this project,
cyborgs and nomadic subjects are companion species that endure.”53 Thus, the posthu­
man is being constituted and affirmed as a hybrid species that combines cyborgs and
nomadic subjects, as nomadic cyborgs.
 JOURNAL OF AESTHETICS AND PHENOMENOLOGY 221

In questioning the current modes of functioning of AI in art, Manovich understands

that AI integration can lead us to limit our thinking about other approaches to aesthetics
in the digital age and the possibility of future reductions in the diversity of cultural
expressions by automation technologies. Like Manovich, Zylinska is concerned with AI
in art, and its involvement in social, political and ecological dimensions, and warns us
about the world of data we live in during “a time when not only do we have access to
more variegated data but also when we ourselves have all become data subjects.”54 As
users or beneficiaries, we need to understand fully the present of AI, in order not to lose
our ability to reimagine a world shared with AI systems.
Critically, the works of Marc Lee, Noah Levenson, Dries Depoorter, Paglen and
McCarthy prompt us to think of the operating systems and functioning of AI technol­
ogies, their storage and sharing systems, the logic of multiple benchmark training sets,
and the dynamics of surveillance. In Conversacube (2010), McCarthy posits a sick society
that seeks to discard any controversy and suppress any interaction that escapes a moral
ideal of perfection, where “ideal” social standards of behaviour are determined by
a normative pattern of consensus or of non-controversy arrived at through AI. In line
with Simondon’s and Hui’s position that the humanization process is linked to the
production process of objects, McCarthy investigates how processes linked to AI can
be understood critically. As Zylinska puts it, this means that “McCarthy raises broader
questions about how ‘individuation’ happens in the age of AI: i.e. how we become human
in relation to technical objects.”55
Paglen’s proposals focus on machine vision, and demonstrate how technological
machines capture images/data, create databases, compare, classify, analyze and share
information socially. The ImageNet Roulette project (2019), led by Paglen and produced
by Leif Ryge, questions the classificatory processes within ImageNet by building an
application that uploads images and reveals how they are labeled, classified and categor­
ized. The work shows that this is not an unbiased process and that it results in stereo­
typing human subjects according to gender, religion, age, activities, and professions
through non-neutral nominal designations. “This focused considerable media attention
on the influential collection’s longtime inclusion of racist and sexist terms”56, with
recurring discriminatory and misogynistic biases. Paglen thus leads us to reflect on
how ImageNet’s taxonomic criteria propagate established human prejudices through
algorithmic reproduction and how AI is programmed to replicate the limiting categor­
izations of language-based conceptual constructs that limit the perception of difference
and alterity.
But criticism cuts both ways, and some aspects of AI in art have been severely
critiqued. For example, Akten’s audio/video installation Deep Meditations: A brief history
of almost everything in 60 minutes, is a one hour abstract film, which seeks to celebrate
life, with images and sounds produced by a deep artificial neural network which
investigates the interaction between humans and non-humans, as a complex ecosystem:
“The work embraces and celebrates the interconnectedness of all human, non−human,
living and non−living things across many scales of time and space – from microbes to
galaxies.”57 However, even though Akten brings this sense of belonging and complexity
between micro- and macrocosms, Zylinska criticizes productions like Akten’s which
naively engage with technological advancement:
222 A. MACHADO OLIVEIRA

Kogan, Tyka, Akten and Klingemann adopt a worryingly uncritical instrumentalism, where
seemingly child-like curiosity is underpinned by the progressivist model of technological
expansion towards some kind of improvement – of accuracy, data sets and, ultimately, art as
we know it. They are thus poster boys for ai art as underwritten by Google in the way they
incarnate the very ideas of progress, innovation and upward trajectory that drive the gung-
ho progressivism of the AI 2.0 era.58

Zylinska points out that there is an attempt to seduce the public through a technological
fascination that always asks for more progress even though ultimately this has a pacifying
and anesthetic effect. Instead of controlling via coercion and prohibition, there is
a control of pleasure and desire where the technological charm produces an illusion of
diversification by altering perceptions, but “without being able to account for differences
that matter – for why and where they matter, when and to whom.”59 AI is permeated by
political issues — for example, whether or not to ignore the biases of neural networks —
and so the political and the algorithmic dovetail with one another in AI.
By altering our sensory apparatus and perceptions, AI art productions raise new
aesthetic and conceptual questions brought to bear on artistic practices: this can be as
innocuous as excluding noises and glitches from the system, or as charged as giving
visibility to certain aesthetic patterns and cognitive modes. By proposing novel sensory
experiences and examining how and why we attribute meaning, AI art can discern
patterns of possibility that can inform the future while reviewing the past for threats
and chances of change in a world permeated by AI. AI art can help us to view afresh the
conceptual and discursive boundaries of human perception, the creation of human and
cultural values, and the recognition of our becoming (with) machines.
From a practice-led research approach, we can perceive the implications of AI within
artistic practices, as well as how these practices have directed new research and produc­
tions on AI by raising questions about how data is configured, stored and shared. In
addition, AI in art enables the discernment of patterns generated from what and how we
think and feel, and facilitates the generation and extrapolation of novelty as future
possibility and the development of new sensibilities.
For Hui, art serves as a way of shaping the sensory and manipulating sensibility
towards a specific method of aesthetic knowing as engaging the non-rational.60 When
reviewing art and technology, Hui examines the varieties of art experience to ask what
these experiences can contribute in terms of heightening sensibility, and to study how
emerging technologies can enhance our senses or improve our capacity to feel. Hui
advocates the need for an Education of Sensibility in order to expand our experiences
beyond the references of calculable and computational rationality through a non-rational
sensibility, which is neither irrational or rational.61
This article has explored the direction AI is offering art productions that, through
human and non-human assemblages, operate under the control of bodies and subjectiv­
ities. We have considered how AI-oriented art practices can help us project futures from
the present, and how they encourage us to open ourselves to speculative imaginings of
possible and implausible futures. These recurring questions in art, philosophy and other
fields of knowledge remain open and prompt us to the future imaginings of what we call
humanity. In order to think the worlds built through emerging technologies, we ask with
the theorists mentioned above, what kind of society do we want to live in, and what roles
will we allow ubiquitous computational technologies, such as AI, to play in our quotidian.
 JOURNAL OF AESTHETICS AND PHENOMENOLOGY 223

Notes
1. Simondon, On the Mode of Existence of Technical Objects.
2. Hui, Art and Cosomotechnics; Hui, ‘On the Limit of Artificial Intelligence’.
3. Manovich, AI Aesthetics.
4. Zylinska, AI Art, Machine Visions and Warped Dreams.
5. Mulder, From Image to Action, 15
6. Oliveira, Pós-digital e Inteligência Artificial: a importância de um paradigma tecno-ético-
estético
7. Simondon, On the Mode of Existence of Technical Objects
8. Haraway & Kunzru Antropologia do Ciborgue; Haraway, “Compoundings”; Braidotti, “All
Too Human: Towards a New Process Ontology”; Braidotti, The Posthuman; Braidotti,
“Posthuman Critical Theory”.
9. Benjamin, Magia e técnica, arte e política: ensaios sobre literatura e história da
cultura
10 Hui, Art and Cosmotechnics.
11. Hui, “On the Limit of Artificial Intelligence”, 341.
12. Haraway & Kunzru Antropologia do Ciborgue.
13. Haraway, “Compoundings”, 119.
14. Haraway, Staying with the Trouble: Making Kin in the Chthulucene.
15. Braidotti, “All Too Human: Towards a New Process Ontology”, 202.
16. Braidotti, The Posthuman.
17. Hui, Recursivity and Contingency, 145.
18. Ibid., 31
19. Ibid., 254.
20. Ibid, 1.
21. Hui, “What Begins After the End of the Enlightenment?”, 8.
22. Hui, Art and Cosmotechnics.
23. Hui, Recursivity and Contingency, 246.
24. As Hui puts it: “Recursivity is not mere mechanical repetition; it is characterized by the
looping movement of returning to itself in order to determine itself, while every movement
is open to contingency, which in turn determines its singularity” (Ibid, 3).
25. Ibid, 341.
26. Hui, Art and Cosmotechnics.
27. Benjamin, Magia e técnica, arte e política: ensaios sobre literatura e história da
cultura
28. Candy, L. Practice Based Research: A Guide, 1.
29. Manovich, AI Aesthetics, 4.
30. Ibid, 18.
31. Ibid, 18.
32. Zylinska, AI Art, Machine Visions and Warped Dreams, 13.
33. Ibid, 13.
34. Ibid, 14.
35. Ibid, 66.
36. Ibid, 70.
37. Agüera y Arcas, B. “Art in the Age of Machine Intelligence.”
38. To name but a few: Thomas Feuerstein, Theo Triantafyllidis, Martine Rothblatt and Lynn
Leeson, Katja Notivskova, Leonel Moura, Guido Segni, Joy Buolamwini, Sougwen Chung,
Daniel Ambrosi, Taryn Southern, Mike Tyka, Pierre Huyghe, Anna Ridler, Gene Kogan,
Stephanie Dinkins, Mimi Onuoha, Wayne McGregor, Tom White, Ross Goodwin, David
Young, Sofia Crespo, Scott Eaton, Ahmed Elgammal, Gauthier Vernier, Ali Nikrang, Mario
Klingemann, Trevor Paglen, Marc Lee, Noah Levenson, Dries Depoorter, Lauren Lee
McCarthy, and Memo Akten.
39. Atken, AIArtists.org.
224 A. MACHADO OLIVEIRA

40. Stocker, Jandl, and Hirsch, The Practice of Art and AI, 27.
41. Ibid, 221.
42. Zylinska, AI Art, Machine Visions and Warped Dreams, 13. “ImageNet is an image database
organized according to the WordNet hierarchy in which each node of the hierarchy is
depicted by hundreds and thousands of images. The project has been instrumental in
advancing computer vision and deep learning research. The data is available for free to
researchers for non-commercial use” (Image-Net).
43. Ibid, 96.
44. Crawford, Atlas of AI, 141.
45. Stocker, Jandl, and Hirsch, The Practice of Art and AI, 218.
46. Zylinska, AI Art, Machine Visions and Warped Dreams, 133.
47. McCarthy, Lauren Lee McCarthy.
48. Parisi, “The alien subject of AI”, 29.
49. Simondon, On the Mode of Existence of Technical Objects.
50. Hui, Art and Cosmotechnics.
51. Hui, Recursivity and Contingency, 244.
52. Hui, “On the Limit of Artificial Intelligence”, 354.
53. Braidotti, “All Too Human: Towards a New Process Ontology”, 207.
54. Zylinska, AI Art, Machine Visions and Warped Dreams, 146.
55. Ibid, 135.
56. Crawford, Atlas of AI, 142.
57. Stocker, Jandl, and Hirsch, The Practice of Art and AI, 246.
58. Zylinska, AI Art, Machine Visions and Warped Dreams, 82.
59. Ibid, 83.
60. Hui, Art and Cosomotechnics.
61. Ibid.

Acknowledgments
This work was supported by the National Council for Scientific and Technological Development/
CNPq/Brazil [Funding Agency #1] and Research Support Foundation of the State of Rio Grande/
FAPERGS/Brazil [Funding Agency #2].

Disclosure Statement
The authors report no conflict of interest.

Notes on contributor
Andreia Machado Oliveira is a multimedia artist, professor and researcher in art, science and
technology on interactive systems, artificial intelligence, the technological image, and collabora­
tion processes. Researcher at the National Council for Scientific and Technological Development -
CNPq/Brazil. Post-doctorate at the City University of Hong Kong, under the supervison of Yuk
Hui, and doctorate at the Federal University of Rio Grande do Sul, and at the Université de
Montreal, under the supervision of Tania Fonseca and Brain Massumi. Currently associate
professor at the Department of Visual Arts and the Graduate Program in Visual Arts, and
founder/director of LabInter/UFSM. https://www.ufsm.br/laboratorios/labinter/

ORCID
Andreia Machado Oliveira http://orcid.org/0000-0002-8582-4441
 JOURNAL OF AESTHETICS AND PHENOMENOLOGY 225

Bibliography
Agüera y Arcas, B. Art in the Age of Machine Intelligence. Arts 6 (2017):18.
Akten, M, AIArtists.org. https://aiartists.org/memo-akten. Accessed 7 October 2022.
Akten, M. An artificial intelligence discussion with visual artist Memo Akten. Downloaded from:
https://www.youtube.com/watch?v=-H-XIVcMY-o. Accessed: 21 January 2022.
Benjamin, W. Magia e técnica, arte e política: ensaios sobre literatura e história da cultura. São
Paulo: Brasiliense, 1994.
Braidotti, R. Posthuman, “All Too Human: Towards a New Process Ontology.” Theory, Culture &
Society, 23 (7–8) (2006): 197–208.
Braidotti, R. The Posthuman. London: Polity, 2013.
Braidotti, R. “Posthuman Critical Theory.” Journal of Posthuman Studies, vol. 1. n. 1 (2017): p. 9–25 .
Broussard, M. Artificial Unintelligence: How Computers Misunderstand the World. Cambridge,
MA: MIT Press, 2018.
Candy, L. Practice Based Research: A Guide. CCS Report 1 (2006):1–19.
Crawford, K. Atlas of AI. New Haven: Yale University Press, 2021.
Haraway, D. & Kunzru, H. Antropologia do Ciborgue. Belo Horizonte: Autêntica, 2000.
Haraway, D. “Compoundings.” In Sensorium: embodied experience, technology, and contemporary
art, edited by C. Jones, Cambridge, MA: The MIT Press, 2016.
Haraway, D. Staying with the Trouble: Making Kin in the Chthulucene. Durham, NH: Duke
University Press Books, 2016.
Hui, Y. Recursivity and Contingency. London: Rowman & Littlefield, 2019.
Hui, Y. “What Begins After the End of the Enlightenment?” E-flux Journal # 96, 2019.
Hui, Y. Art and Cosmotechnics. Minneapolis, MN: e-flux, 2021.
Hui, Y. “On the Limit of Artificial Intelligence.” Philosophy Today, Volume 65, Issue 2 (Spring
2021): 339–357.
Image-Net, https://www.image-net.org/index.php. Accessed 7 October 2022.
McCarthy, L. Lauren Lee McCarthy. Downloaded from: https://lauren-mccarthy.com/conversa
cube. Accessed: 10 Jan. 2022.
Manovich, L. AI Aesthetics. Moscow: Strelka Press. Kindle edition, 2018.
Mazzone, M.; Elgammal, A. Art, Creativity, and the Potential of Artificial Intelligence. Arts,
8 (2019): 26.
Mulder, A. From Image to Interaction. Rotterdam: V2_Publishing, 2010.
Oliveira, A.M. “Pós-digital e Inteligência Artificial: a importância de um paradigma tecno-ético-
estético.” In: C. Rocha et al. (Orgs). Anais do VII Simpósio Internacional de Inovação em Mídias
Interativas. HUB Eventos 2020. São Paulo: Media Lab/BR, PUC-SP.
Paglen, T. This Artist Shows Us How Computers See the World. Downloaded from: https://video.
vice.com/en_us/video/this-artist-shows-us-how-computers-see-the-world
/5976485b68e02bb371e2301d. Accessed: 16 Feb. 2022.
Parisi, L. “The alien subject of AI.” Subjectivity, vol. 12 (2019): 27–48.
Simondon, G. On the Mode of Existence of Technical Objects, Translated by, C. Malaspina,
J. Rogove. Minneapolis: Univocal Publishing, 2017.
Stocker, G., Jandl, M., and Hirsch, A. J. The Practice of Art and AI. Linz, Ars Electronica Linz
GmbH & Co KG., 2021.
Zylinska, J. AI Art, Machine Visions and Warped Dreams. London: Open Humanities Press, 2020.