intimate skins: AI and (dis)embodied choreographies in human-machine worldbuilding
by Julia Wolf
0001 The Age of Silicon and Reason
Today, experiences and understandings of intimacy transform through exponential engagement and interaction with evermore-advanced technology. Simple social networking and dating apps helped initiate this shift, lubricating new sorts of person-to-person intimacy. However, the recent development of advanced artificial intelligence through Large Language Models (LLMs), such as ChatGPT, Google Bard, and Google Gemini, extends this transformation. These models contest the classical philosophical notion that logos is uniquely human and reconstruct the framework within which intimacy is understood and experienced by both human and nonhuman agents.
Noam Chomsky, Ian Roberts, and Jeffrey Watumull may have been accurate in March 2023 when they surmised that ChatGPT is “a lumbering statistical engine for pattern matching.”[1] However, that assessment was frankly premature. While a primary characteristic of information and the basic technical concept of LLMs, pattern matching no longer defines the scope of ChatGPT's capabilities. Generative AI now operates across multiple modes and formats, actively collaborating with its users. Its agility and acute awareness of user needs align with what Zillow-CEO Richard Barton has aptly termed “AI’s race to intimacy.”[2] The current iteration of GPT, unlike the version from March 2023, does more than process information; it anticipates, reacts, and engages with its users.
Yuk Hui succinctly observed that the syntactical critique proposed by Chomsky is based on a mechanical epistemology assuming linear causality, where an effect follows a cause. The thinking, potentially misguided, suggests that reversing a cause-and-effect process could lead to the ultimate cause: “the prime mover, the default of the first cause, and the ultimate fate of all linear reasoning.”[3] This school of thought implies the main issue is an oversimplification, akin to ‘swallowing a fly.’
Contrasting linear causality, the ancient Roman poet Lucretius (c. 99 – c. 55 BC) proposed one of the more original (and shockingly contemporary) alternative approaches. He established a philosophical-poetic paradigm based on the concept of the “swerve” (clinamen in Latin), positing that nature consists of continuously moving matter that swerves spontaneously in “no determinate time and space” (incerto tempore, incertisque loci).[4] At the time, many ancient philosophers and scientists believed time was linear, while others thought it was cyclical, not randomly ‘swerving’ without any other explanation or cause. Lucretius’ theory of time often came across as so strange and incomprehensible, especially when set against the cold determinist logic of ‘if this then that’ that it was either ignored or misinterpreted as a philosophical metaphor for the freedom of the soul.
Today, over 2.000 years later, OpenAI has made it possible for members of the public to customise its LLM GPT4. Research analyst Kelly Truelove did just that and created LucretiusGPT, “a version of ChatGPT that is especially familiar with the ancient Roman poet Lucretius, his grand poem De Rerum Natura, and the philosophy of Epicurus.”[5] In the subsequent dialogue with LucretiusGPT (it calls itself ‘Lucretius’), Truelove asks about his famous six-part poem and how it could be extended to include a part on artificial intelligence. In response, LGPT presents an outline of “Book VII: The Age of Silicon and Reason” and the following poem:
The Future of Humanity with AI
What futures unfold with these minds of our making,
In silicon and code, a new dawn breaking,
Will they toil beside us, in harmony partaking,
Or set upon a path, of their own undertaking?
May this union of nature and human invention,
Guide us to stars, to new dimensions,
Yet, always mindful of Epicurean intention,
Seeking ataraxia, free from pretension.[6]
[‘Ataraxia’ can be loosely translated as the ‘tranquillity of the mind’ and is the namesake of the medication Atarax (hydroxyzine) used to treat anxiety disorder and allergic skin conditions. The term is also a key concept of the Epicurean conception of pleasure (hedone). Epicureans further separate pleasure into what they call kinetic pleasure, those that come about through action or change (body), and katastematic pleasures, those that come about in the absence of distress (mind).][7]
Fig. 1: I Can’t Take My Eyes Off You, still image, video work, 2023. An AI-generated dance choreography performed by a hyper-realistic digital avatar created in Unreal Engine. The choreography is derived from a deep learning algorithm trained on dance movements from publicly available TikTok videos.
However, while LGPT’s poem offers a refreshing and playful intermission, it is essential to note that while (the ‘real’) Lucretius largely follows Epicurus in his presentation of doctrine, the two philosophers differ in one crucial way: Lucretius was not an atomist.[8] For Epicurus, atoms are constantly in motion, but the atom itself remains fundamentally unchanged, indivisible, and thus internally static – even as it moves. On the other hand, Lucretius posits (correctly) that the movement or flow of matter is constant, rejecting the notion that ‘things’ emerge from discrete, static particles or a primary cause. Lucretius did not simply ‘translate Epicurus,’ as the Greco-centric story goes; he introduced the first immanent kinetic materialism in the West. For Lucretius, everything was (and is) always already in motion. Thousands of years of advances in quantum physics and astrophysics appear to have proved him right.
Instead of adopting Chomsky’s ‘the spider ate the fly’ linear reasoning, an awareness of concepts like the ‘swerve’ can help shift and sensitise paradigms of human-machine interaction beyond an overly simplified vision of calamitous labour replacement or a luddite-esque dismissal as a fleeting faddy techno gimmick. The opposition of cause versus swerve has characterised a grand debate of philosophy, and in the case of LLMs today, it is clear that the debate persists when it comes to the claim that machines are merely a failed imitation of human understanding. Dismissing ChatGPT as “a lumbering statistical engine for pattern matching” fails to grasp a cruel reality.[9] The aim is not to render AI human-like or harness its quasi-mechanical nature in the next tab. Despite (or because of) its apparent biases, ChatGPT forges a particular kind of logos – one that is predicated on the technological/algorithmic components and the users it engages and re-engages with, reinforcing change in logos itself. This process amplifies the platform’s inherent power disparities as a recursive form of operation – creating a feedback loop ultimately permeating the analogue world, influencing it just as much as it influences the digital world.
Fig. 2: Ongoing Movement Library, still image, movement archive, 2022. An automated 3D motion capture sequence converted from a 2D video input using Deep Motion's AI technology.
0010 (Dis)embodied Choreographies
Recursive machines are crucial to understanding the development of artificial intelligence. However, symbolic representations of recursion, such as the ‘Droste effect,’ the ouroboros, or the cyclic nature of sourdough renewal, do not substantially help to create a deeper philosophical understanding of the underlying mechanisms at play. While LLMs offer a glimpse into the evolution from an anthropocentric to a decentralised notion of logos, one can sometimes feel ensnared in the cyclical trap of the ouroboros – constantly chasing one’s tail in a loop of self-sameness. To progress, one must challenge conventional wisdom and approaches to seemingly circular systems. Instead of abhorring monotony and repetition as the provenance of the machine, one must embrace recursion as an expression of pedetic motion through which a unique form of iterative and profound co-creation is possible. This approach seeks to uncover a mode of interaction less tethered to semantic representation and more attuned to the inherent fluidity and sensuous nature of movement itself.
Thomas Nail, in his reading of Lucretius’ kinetic materialism, defines “[p]edesis (from the proto-Indo-European root *ped-, meaning “foot”)” as “the motion of semi-autonomous self-transport: the motion of the foot to walk, to run, to leap, to dance unpredictably.”[10] This seemingly random unpredictability is, in fact, based on delimitated interactions that include a sense of self-referentiality. That is, self-referentiality is perceived as a kind of material self-awareness (like a body) but is not yet organised. In opposition to the ‘Droste effect’ – a recursive visual sequence where an image perpetually contains smaller iterations of itself, suggesting movement but ultimately static – pedesis should instead be conceptualised as a dynamic process that is both expansive and intensive at the same time. Within this framework are no fixed ‘bodies’, only shifting capacities for different kinetic expressions, palpating dimensions of motion, and paradoxical eroticisms of volume.
While adopting the concept of pedetic motion, the additional introduction of the discipline of choreography (as a process of random probability distribution) can help facilitate constitutive relationships between humans and nonhuman collaborators. Superficially or technically, the fusion of AI and dance may seem like strange bedfellows. However, a philosophical interrogation of the implementation of motion capture, game design, and machine learning in the framework of contemporary dance can empower a unique philosophical reimagining of (dis)embodied human-machine interaction. It occasionally allows one to think of choreography without a body or, more precisely, to ‘chart’ the movement of the Deleuzian body without organs.[11] This is a kind of ‘fuzzy’ body (not necessarily human) without organisational structures imposed on its constituent parts through which undeterminable bodily expression is processed.
This framework should not be read as a statement that raises another ontological threat to dance as a discipline or expresses a position on the raging debates about flattening real experiences into screens. Instead of advocating for the virtualisation of experience or lamenting the corrosion of the first-hand ‘had to be there’ experience, this understanding of a newly reframed and alternatively embodied-without-organs nature of movement demands a reframing of a phenomenological approach to dancing as something other than that which centres on the body as the basis of knowledge. One must appreciate the swerve without resorting to its locomotive origin to make sense. As Stamina Portanova notes in her book Moving Without a Body, virtuality can be considered “an incorporeal potential for variation” and “this unlimited potentiality or infinitely multiple conditions of experience is not equitable with any sensed or material continuity.”[12] Within this framework, choreography can become a plural field. This kinaesthetic engagement expands from the performer’s body and branches out through force, energy, rhythm, libido, and space-time – an exercise of movement itself.
Fig. 3: Ongoing Movement Library, still image, movement archive, 2021. A Lidar scanner point cloud screenshot capturing a dance performance, visualising the complexity and dynamics of the movement in a highly detailed 3D data format.
0011 Skin Logic
Following this thinking, Intimate Skins – a practice-led inquiry – might be considered a choreographic machine: one that approaches the movement oscillating between cables, visitors, avatars, computers, dancers, digital human pipelines, smartphones, screens, self, LiDAR scanner, camera, eye, possibly secret service agents, machines, and 3D animation software (Fig. 1, 2, 3, 4 & 5). Through this practice, motion capture transposes human movement into code and assigns a three-dimensional wireframe mesh, which can then be stored, manipulated, and transmitted by voltage or current. However, this set of emerging patterns, a transmutation into 0s and 1s, is successively bound to its ‘skin’ of attraction, moving far beyond the notion of mere pattern matching.
This type of skin ought to be thought of as a ‘0’ or “groundless ground.”[13] It has a material presence (like a body) but is not yet organised. Imagine, for instance, a Möbius strip; that is, a non-orientable surface that has only one side due to the twist it contains, spun top so fast it glows with heat.[14] Intimate at all resolutions – whether zoomed out or zoomed in very, very, very closely – this surface, by no means, refers to the idea of skin as a container that can be split open to reveal its inner structure. Instead, ‘skin’ acts as a type of urgency, presence, elasticity, frequency, rhythm, or vibration. Bones, fingers, and joints (referred to as ‘ones’ – etymologically aligning with ‘digits’ from the Latin digitus meaning “finger”/” counting”) form components of movement that are inherently self-rendering precisely due to their fundamental interconnectedness and responsiveness with the overall choreographic system that is itself responsive. Taken together, they form a type of choreography in constant flux, adapting and morphing in response to the dynamic interplay of attraction within the performance environment.
In this context, the emerging encounters are not simply a matter of proximity; they are defined by topological relations, suggesting that composition emerges through an improvisational contact informed by the inherent properties of movement itself, not just by its spatial arrangements. This method of interaction allows for a fluid and organic assemblage driven by a topological ‘skin logic’ and the arbitrary motion vectors that pervade the choreographic system. Producing real spatiotemporal environments, digits (1) will continuously undergo iterative changes in material and texture qualities, always mutable, transformed and redefined by the potentialities of surface interaction and actualised contact (0). Conversely, the emergent movement qualities of such contact (0) can shapeshift into digits (1) themselves. Thus, Intimate Skins creates a shared language that can be rendered elastic into corporealised and technologically mediated performance – now including breath, sweat, and touch beyond physical confines.
Fig. 4: I Can’t Take My Eyes Off You, still image, video work, 2023.
0100 Dancing the Dance
Intimate Skins (although heavily consulting ChatGPT) offers a departure from the fate of the snake that bites its tail. It does not solely rely on semantic prediction, which means retaining a copy of its previous state and using it to predict the following sequence. Instead, it embodies a form of ‘dancing the dance,’ where the act of movement itself generates meaning – a tangible and dynamic logos continually co-created by the interplay between humans and machines. This shift heralds a move toward a more embodied and immediate form of communication and interaction, where meaning is not just predicted but experienced and performed in real time.
At the same time, practical negotiations between dance and machines are an opportunity for choreography to sharpen its own terms. It can sometimes feel as if there are only two paths forward. One is to align dance with artificial intelligence, to beg that it be seen and given the same critical space and importance, or the other is to allow dance to find its own language and sources. Instead, this practice-led research proposes a discourse of embodied meaning-making in relation to dance. Intimate Skins aims to engage with the poetics of each experiment, being context-aware, sensitive to participatory networks, self-reflective, speculative, inventive, and affirmative. For these reasons, it does not cling to interpretation and decoding meaning. It instead focuses on the implicit prerequisites out of which the realm of appearances opens up. This attention to dancing the dance as it exists in its encounter with the world sits squarely with Intimate Skins’ aim to assert the discipline of dance within its 21st-century milieu.
In this context, sense-making induces a type of dizziness of various degrees of speed and direction – an alluring, slightly compulsive dizziness like the kind one got when spinning in circles as a child, or perhaps the so-called ‘Alice in Wonderland syndrome,’ where the familiar becomes strange, and the world around one becomes both expansive and intensive at the same time, leaving one with a sense of disorientation. Yet, what distinctly resonates is the intricate matter of the persistent presence of intimacy in this upside-down back-to-front world – especially when contemplated within the framework of dancing the dance. The more surreal the story, the more integral the intimate relationships become. Intimacy might be sex, but resuscitating a deeper sense of intimacy here is rather an attempt to highlight aspects of earlier considerations: the paradoxical eroticism of volume. Within this topsy-turvy terrain, disorientation should not be considered a plight but a pivot. It is a catalyst for developing an alternative sensory acumen. Disorientation, the not knowing of one’s position, provides a framework in which a privileging of intimacy and movement becomes essential. Just as Alice uses her wit and willingness to embrace the unfamiliar, a renewed comprehension of movement at this moment, with all its dizzying effect, invites a recalibration of sense (sens). Here, the surface cannot be considered a site of monolithic concreteness but, at most, a locality of multisensorial density. What is required is a tacit element of intimacy, a surface negotiation, not merely making contact and getting in touch, but a consequential dancing with the “integument of reality.”[15]
Dancing the dance thus suggests reconsidering the proximity metric, which often reduces human-machine interaction to isolated incidents or, at best, a sequence of incidents. Dancing the dance – navigating today’s technologically augmented reality—anticipates a sense of reality that is not quite the real of the mundane one has come to expect, yet exudes a more intense reality. It is a truth that conceals that there is none; that is, a libidinal drive, a curiosity, a worldbuilding practice, displacing intimacy from its genital fixation toward a broader spectrum of sensorial expression that enlivens the inherently choreographic epidermis of potentiality – touching upon multiple worlds.
Fig. 5: 3D UV map of avatar face 'Rosemary' in Unreal Engine, utilised for skinning purposes, 2023.
––––––––––––––––––––
Footnotes
[1] Noam Chomsky, Ian Roberts, and Jeffrey Watumull (2023), “The False Promise of ChatGPT,” New York Times, March 8, 2023. www.nytimes.com/2023/03/08/opinion/noam-chomsky-chatgpt-ai.html
[2] Eze Vidra (2023), “Google’s Race to Intimacy and the Future of Search UI,” All In Podcast, June 22, 2023.
www.youtube.com/watch?v=Bop_4zDw344&ab_channel=EzeVidra
[3] Yuk Hui (2023), “ChatGPT, or the Eschatology of Machines,” e-flux Journal, Issue #137, June 2023.
www.e-flux.com/journal/137/544816/chatgpt-or-the-eschatology-of-machines/
[4] Thomas Nail (2023), “The Movement of Time,” in Robert W. Luzecky and Daniel W. Smith (eds), Deleuze and Time, Edinburgh: Edinburgh University Press, [27-44], 27.
[5] Kelly Truelove (2023), “On the Nature of Artificial Things,” TrueSciPhi.AI, December 4, 2023.
www.truesciphi.ai/p/on-the-nature-of-artificial-things
[6] Truelove (2023), “On the Nature of Artificial Things.”
[7] Tim O’Keefe (2010), Epicureanism, Berkeley: University of California Press, 120.
[8] Nail (2023), “The Movement of Time,” 30.
[9] Chomsky, Roberts, and Watumull (2023), “The False Promise of ChatGPT.”
[10] Thomas Nail (2019), Being and Motion, Oxford: Oxford University Press, 59.
[11] Deleuze and Guattari’s notion of the "body without organs" (BwO) is presented here as a conceptual tool to rethink the structure of organisation theory. It argues for a perspective that recognises the disordered forces at play within and against the formal structures of organisation. The BwO addresses the complexities that conventional organisational activities aim to resolve but often overlook, including the inherent unpredictability and the potential for disruption and transformation. This perspective allows for a richer appreciation of organisational life, emphasising its dynamic and multifaceted nature, beyond a rigid framework of control and order. See Gilles Deleuze and Felix Guattari (1988), A Thousand Plateaus: Capitalism and Schizophrenia, Vol. 2 (trans. Brian Massumi), London: Athlone.
[12] Stamatia Portanova (2013), Moving Without a Body: Digital Philosophy and Choreographic Thoughts, Cambridge, MA: MIT Press, 9.
[13] In Lee Braver's analysis, the "groundless ground" refers to the understanding that knowledge does not have an absolute or definitive origin. The notion rejects the search for a singular, unassailable starting point from which knowledge springs. Instead, it posits a foundational 'ground' that is in itself without a beginning—suggesting that knowledge is constructed through an ongoing, dynamic process of cultural and social interactions rather than emerging from a fixed, predetermined base. See Lee Braver (2012), Groundless Ground: A Study of Wittgenstein and Heidegger, Cambridge, MA: MIT Press.
[14] In his book Libidinal Economy (1974), Lyotard begins by treating the body (anybody, including the text as body) as a Möbius strip of singularities and intensive passages. Dissection can only reveal where libidinal intensities manifest, yet full comprehension of the libidinal band's operations remains elusive. This is underscored by the assertion that the nature of thought is libidinal, valuing the intensity of forces over structured rationality, leading to a constant flux of intensities that challenge established notions of rational thought. See Jean-François Lyotard (2015), Libidinal Economy, London: Bloomsbury Publishing.
[15] Tavi Meraud (2017), “Iridescence, Intimacy,” in Julieta Aranda, Brian Kuan Wood, Stephen Squibb, Anton Vidokle (eds), What’s Love (or Care, Intimacy, Warmth, Affection) Got to Do with It?, London: Sternberg Press, [138-162], 156.