The present flickers. Like a permanently dislocated shoulder, it snaps between the promise of unlimited freedom and paralysing anxiety and uncertainty. Perhaps a sense of dislocation, without hope of full resolution, has accompanied every era in its immediate experience. It only appears stable in the rear-view mirror. Today’s neuro-totalitarianism 1 seems eager for the dreaded tomorrow. Data flows then make this experience of multiplicity resonate even more intensely, trapping us in information storms with no apparent outside. Reality then seems to be experienced as an ongoing calibration of the meaning-making banality of everyday acts. Large-scale history, in endless loops, inscribes itself upon us just as it slips through our fingers.

Neuro-totalitarianism, a term coined by Franco ‘Bifo’ Berardi, refers to a new form of power that operates not through prohibitions, censorship or physical repression, but through the constant pressure of information, signs and images on the nervous system. Franco ‘Bifo’ BERARDI, Futurability: The Age of Impotence and the Horizon of Possibility, London: Verso 2017.

Andreas J. Hirsch describes artificial intelligence as a phantom – a spectre haunting the world, gradually permeating every layer of the everyday and becoming visible only partially. “It comes in the guise of the epitome of progress, promising increased efficiency and profit to some, absolute control and power to certain others, and commodities to the masses. Some see it as the harbinger of doom for humankind, a Pandora’s box of self-destruction now opened and impossible to close again” 2. The motif of a spectre in between the subject of desire and dystopian force illustrates and (re)produces a narrative that has gained considerable attention recently. This narrative frames the debate and our relationship with artificial intelligence based on a reductive duality of good and evil.

Andreas J. HIRSCH, „Five Preliminary Notes on the Practice of AI and Art“, in: Gerfried STOCKER – Markus JANDL – Andreas J. HIRSCH (eds.), The Practice of Art and AI, Linz: Ars Electronica 2021, p. 12.

How can we break free from the black-and-white tentacles of technophilia and technophobia when the media landscape is constantly ruffled by apocalyptic predictions interspersed with tech bros’ fetishising visions? I will now attempt to undertake a journey beneath the surface of these oversimplifications and build upon the tradition of thought striving to critically disentangle (artificial) intelligent systems from its anthropocentric foundation through various imaginative practices: visual culture and artistic practice that actively engage with the mechanisms of artificial intelligence. This allows us to move beyond the binary framework towards a more productive conception of technology as an infrastructure. In that way we can re-evaluate the perception and connection between human and machine and the emergence of alternative forms of these relationships.

A crisis of intelligence?

When we speak of artificial intelligence in relation to the visual field and art, we are referring primarily to a type of generative machine learning that operates on the principle of extensive datasets and the text conversion to visual media. The image streams of social media are currently heavily populated by these AI-generated scenes. In an accelerated flow of visual production we encounter cute animal videos and yellowish Ghibli-style figures, but also political phantasmagoria often from far-right political personalities (let’s name, for example, the video Trump Gaza shared by Trump, and in the Czech context, infamous racist SPD billboards) only to be replaced in a split second.

Hito Steyerl refers to generative models based on datasets as “mean images”. She interprets the word “mean” in multiple senses: as evil (moral), as average (statistical), as low (class), as meaningful (semantic), and as a means or instrument (economic). 3 Models such as DALL-E, Midjourney and others enable a transition from classical representation, which requires an existing object, to a statistical average calculated from datasets. In Steyerl’s words, this leads to “products of data populism”. 4 The never-ending remixing and recycling of whether human (first generation) or machine-generated (second generation) production by artificial intelligence leads to the homogenisation of significant parts of visual culture and to a crisis of the perception of intelligence and creativity.

Hito STEYERL, „Mean Images“, New Left Review, vol. II, 2023, No. 140/141, https://newleftreview.org/issues/ii140/articles/hito-steyerl-mean-images (accessed 5 March 2026).

Ibidem.

Instead of succumbing to dystopian scenarios, however, this crisis can be productively grasped as an opportunity to reevaluate the very concept of these qualities – which we still exclusively associate with humans – and to disrupt anthropocentric logic, as I indicated in the introduction to this text. This can be achieved if, together with N. Katherine Hayles, 5 we reformulate poorly posed questions such as “Can a machine be creative?”; “Will it relegate us to the role of mere curators and/or to the entirely passive position of spectators?” Instead, we might delve into the nature of machines’ visual modes of communication and pose different questions about it. For example: on what basis do we evaluate a being as intelligent and creative? What escapes us within this framework? And how might we cultivate other forms of intelligence that transcend the human scale?

N. Katherine HAYLES, „Inside the Mind of an AI: Materiality and the Crisis of Representation“, New Literary History, vol. 53 & 54, 2022 & 2023, No. 1, pp. 635–666.

Generative AI scenes unfold new forms of image, visuality and vision based on computational reduction and prediction. The aforementioned content is commonly interpreted in terms of its lack of originality, algorithmic falseness, plagiarism and the glossy spectacularity of late capitalism that distorts our ability to think critically. In this context, Joanna Zylinska speaks directly of “a glorified version of Candy Crush that seductively maims our bodies and brains”. 6 This platform art is also often perceived as the primary representative and product of the visual culture of artificial intelligence. Yet underneath it lies a rich landscape of artistic tendencies that may borrow this aesthetic, deconstruct it and completely distance themselves from it, albeit by using their own mechanisms.

Joanna ZYLINSKA, AI Art: Machine Visions and Warped Dreams, London: Open Humanities Press 2020, p. 76.

Art in the Plurality of Datasets

Me, myself and (a)I from 2025 presents a series of digital images and augmented reality. In this work, Martina Menegon had artificial intelligence trained on her own dataset composed of her selfies. The resulting visual component combines the fleeting experience of the subject, which multiplies across platforms, with the slightly distorted aesthetic of generatively produced forms.

Working with datasets is commonly viewed through the prism of absence, a kind of deficiency that can be understood more generally (the reduction of the world’s complexity) or understood more specifically in the case of the deliberate absence of various experiences (racial and gender biases; the invisibility of those who label and curate dataset content) and of natural resources (high water consumption and regional scarcity). A culture of datasets with power implications embedded in their structures then resonates in artistic efforts to decolonise and “queer” the technology by pointing out absent bodies and subsequently expanding the horizons of AI through alternative datasets. However, eliminating racial bias by merely adding images of minorities also leads to more effective surveillance by the repressive apparatus. 7 Given the current situation, in which OpenAI (which owns ChatGPT and DALL-E) has entered into an agreement with the U.S. Department of Defense to provide artificial intelligence technologies for classified U.S. military networks, 8 it is essential to bear in mind that a more inclusive dataset can simultaneously lead to intensified discrimination.

STEYERL, „Mean Images“.

Hadas GOLD, “OpenAI strikes deal with Pentagon hours after Trump admin bans Anthropic”, CNN Business, 28 February 2026, https://edition.cnn.com/2026/02/27/tech/openai-pentagon-deal-ai-systems (accessed 8 March 2026).

In her work, Menegon addresses machine vision, which corresponds to data structures and the remaking of the body through statistics. Artificial intelligence contributes here to forming of the subjectivity, which is not merely represented but rather continuously generated at the boundary between a biological entity and algorithm. However, the intimacy of a dataset conceived in this way can also be limiting, as it delineates narrow boundaries of the worlds it constructs and reflects. Artificial intelligence can also be perceived as a co-producer of collective memory, where the predicted average generates the user’s profile through a shared experience, as in the case of Alexey Yurenev’s AI-generated war photography, where the images represent a synthetic, collective memory of war. 9 Yurenev utilised a monumental dataset containing historical photographs from World War II. The resulting synthesis then emerges from the collaboration between machine and human in an unprecedented visual restaging of the war.

Bogna KONIOR, „War in the Age of Infinite Evidence: On AI-Generated War Photography“, e-flux Journal, 2026, No. 160, https://www.e-flux.com/journal/160/6776831/war-in-the-age-of-infinite-evidence-on-ai-generated-war-photography (accessed 9 March 2026).

Martina Menegon’s digital images, in their form, reference an older type of generative model (GAN). 10 They are therefore not so much characterised by the current aesthetic of diffusion models which generate images from textual prompts and, as noted above, are often regarded as the dominant form of today’s generative visual production. Contemporary art, however, may draw on the aesthetics of platform slop, the slippery nature of the viral storm in whose eye we find ourselves, and play with the binary between technological utopia and doomism. We can see this strategy in Daniel Felstead and Jenn Leung’s filmic essays, in which we are confronted with the tension between the glossy surface of new technologies and the shattering form of reality such technologies produce.

Generative adversarial networks represent a type of artificial intelligence that employs two neural networks (a generator and a discriminator) which are interconnected and compete with each other. The generator produces realistic data, while the discriminator evaluates it and attempts to determine whether the output is real or artificially generated.

A Landscape of Expectations

Zylinska argues that the primary contribution of art that consciously engages with artificial intelligence in its “ability to redraft the conceptual and discursive boundaries of human perception, human value and human cultural practice, while drawing us as its human recipients into the recognition of our becoming (with) machines.” 11 Generative AI thus functions in the world not only as a technological apparatus but as a socio-political actor woven into a broader economic, political, social and cultural network. In this respect, for Zylinska it does not embody an apocalyptic horseman, but rather represents an expansion of possibilities for developing different types of creativity and intelligence, as well as new approaches to them, where creativity and intelligence emerge through interaction, in hybrid relationships and situations between humans and machines. Artistic practice can de-instrumentalise technology, explore these in-between situations, point out its limits, and speculate on its alternative forms and modes of existence, without necessarily having to stand in opposition to technology.

ZYLINSKA, AI Art, p. 142.

However, I do not wish to downplay the technology’s negative impacts. We have not mentioned yet its unsustainably extractive nature: it demands the world be transformed into raw material for extraction and commodification – a process rooted in its origins, control and utilisation within the logic of late capitalism (or perhaps techno-feudalism?). What has to be also considered is the production of deep fakes, the truly banal, sedative content of the consumer spectacle, mass surveillance, parasitic appropriation of artworks, and the erosive impact on individuals’ mental health (so-called “AI psychosis”). For these reasons, art needs to engage with the technology critically – not merely as a neutral instrument but as a distinct actor situated within a broad network of material-discursive relations. In this respect criticality concerning creativity and intelligence within a distributed network of actors, materials and relationships need not apply exclusively to AI art and new media art; it is an extension of critical thinking about the technology applied already to more traditional media.

In the introduction to this essay, I discussed the present as straddling the seemingly open field of unlimited possibilities and paralysing anxiety within a swirling storm. It has been seventeen years since Mark Fisher diagnosed what he called a crisis of imagination. 12 With U.S. President Donald Trump’s second term and the overall geopolitical events, the news are flooded with the discussion about the end of the world order as we know it. Can the present once again become a landscape of expectation? One hope lies in the ability to tell different stories and collectively shape a new cosmopolitics that liberates technology from the neoliberal framework of extraction and surveillance. The answer, then, is not technology itself, but a situated and ethical way of using it and imagining it. In Zylinska’s words: “one of the most creative – and most needed – ways in which artists can use AI is by telling better stories about AI, while also imagining better ways of living with AI.” 13

Mark FISHER, Capitalist Realism: Is There No Alternative?, Winchester: Zero Books 2009.

ZYLINSKA, AI Art, p. 31.

Denisa Michalinová is a PhD researcher at the Academy of Fine Arts in Prague and focuses on the theory of contemporary art, corporeality and new technologies.