Vol 24 (2025) Aesthetics of Biomachine
Contagious Life: Clones, Deadbots, Digital Twins
Caroline Bassett
University of Cambridge
AI recreations offer a 96.5% similarity to the original person.1
Preface
The internet has long hosted the materials of the dead. Now it hosts the dead themselves, not only as archival traces (as stored ‘old media’ – documents, photographs, stories), but in the form of persons, things, bots or deadbots, all being given some semblance of agency or even of life. At any rate some of these forms suggest life, or claim it, or claim some intimate connection to it – coming close enough to touch it, perhaps; and desired or not, touch brings with it fears of contagion. But what kind of life is this? What fuels it, what guarantees it, what crossings and fusings does it entertain with or require from ‘standard’ biomachines – we ourselves, as technologised humans in the 21st century (Thacker, 2003)? Do these forms constitute a new mode of, or a new engagement between, technologies and the flesh, a new biomachinic contract? Or do they simply simulate life, ultimately remaining safely on the far side of the ditch dividing the algorithmic from the organic, or from what Beatrice Fazi (2019) terms the discrete and the continuous? If these are algorithmic ghosts, are they honest ghosts? And if they constitute or exemplify a new form of life – and in doing so also challenge death – is this an honest aesthetic? What do its promises amount to? This essay focuses on an older formation, and on some of the emerging forms the digital afterlife takes – forms chosen because they can help to articulate a new aesthetic and perhaps point to how it may be adjudged.
It begins in a valley far away, in another century.
Yea, Though I Walk
In the mid- to late 20th century, Masahiro Mori, a Japanese roboticist who was also a Buddhist, speculated on the existence of an uncanny valley, a zone of indeterminacy that might be arrived at on the way to the ‘perfection’ of the artificial hominoid. This valley, he said, should be avoided; designers should learn to take different paths. The valley floor, the dip at the bottom of what Mori laid out as a U-shaped diagram, indicates the moment the desired simulation of the human form closely approaches its object but fails in some aspect, thereby producing a change in sensation: ‘one tiny design change’ and the switch is from empathy and attraction to ‘fear and revulsion’ (Mori, 2012). This is a ‘descent into eeriness’, and the title of Mori’s paper, ‘Bukimi no tani genshō’, directly translates as ‘valley of eeriness’. Later, in 1978, it was renamed ‘the uncanny valley’ by the curator Jasia Reichardt in Robots: Fact, Fiction, and Prediction (Kageki, 2025). ‘Bukimi no tani genshō’ was an explicit response to the rapidly developing field of robotics, but the questions it raises about simulation, imitation and the forms they might take relate to multiple other areas where the increasingly ‘perfect’ simulation of human capabilities, which seems to threaten to bring humans and machines into very intimate relations, produces a parallel unease. The automation of language, itself often taken as indicating if not life then intelligence and perhaps non-fungibility, constitutes a contemporary iteration of this. What might be termed an automation anxiety/avidity (Bassett, 2022; Bassett & Roberts, 2019), evident in both public discourse and more specialist spheres, produces a feeling of revulsion but also attraction, and one that cycles: to take one example, if the Institute of Contemporary Art’s first computer art show, curated by Reichardt in 1968,2 was all about cybernetic ‘serendipity’, the follow-up, in 1994, was ‘seduced and abandoned’ (Bassett, 2025).3
Under discussion here is a new moment of unease arising around new kinds of proximity developing between artificial and human forms, enabled by artificial intelligence (AI) and generative algorithms. Not only proximity but contact is explored, and a key term to think about this is contagion – defined as that which might occur when things (and maybe humans) come very close together, bringing into prospect, or perhaps enacting, an exchange of something variously described as vitality,4 spirit or essence, the results of which might redraw the normative boundaries of the fleshed individual and/or give formerly inert objects a form of independent or quasi-independent ‘life’ – or at least its seeming.
One of the issues to be explored here is the limits and possibilities of what ‘seems’. When we look at a deadbot, is it only a ‘seeming likeness’, entirely distinct from what it seems like, or is there a blurring between what seems and what is (maybe between simile and metaphor), enabled in various ways – that is, is this actually one of the ways in which contagion, exchange, happens? At any rate, contagion here constitutes a relation where separation is insisted on and also collapsed. A parallel is Seth Giddings’s (2024) account of toys, which insists on their integral relation and complex links and pathways to the worlds and things they are often assumed only to model. It is striking that Giddings focuses not on a relationship bounded by simile or metaphor, but on one that also entails synecdoche, the part representing the whole.
For Fazi (2019), the incommensurability of sensation and calculation, the continuous and the discrete, with humans on one side and the computationally produced on the other, means contagion doesn’t occur. It is this separateness that also contributes to what she sees as a ‘deadlock’ in digital aesthetics, since it tends to be assumed that therefore the computational cannot be inaugural or creative. Her way out of the deadlock is to identify an indeterminacy in the computational that might constitute the basis of what may be termed an alien aesthetics.
My wager, also exploring the ditch between the discrete and continuous, technology and culture, life and non-life, is that a new ‘new aesthetic’ may be identified, and it emerges not on one side or the other of the dividing ditch, but at its bottom, where things pool and mix: these days we live in the valley. Accounting for this new aesthetic, or for what this form of culture and life might inhere in, demands that we neither fall into the machine (Fazi’s desire to identify a generative aesthetic in the computational alone) nor reverse into the human (where the anthropomorphic becomes the locus for the measure of what might constitute the new, a move that also tends to prioritise the ethnographic as a methodology). The point is to think about how they connect. However, this does not, in my view, boil down to or become a simple matter of prosthesis, certainly not where this amounts to an adherence to the prosthetic sublime with its (dream of a) seamless and painless extension of the human, an assumption of the machine. As Jonathan Sterne (2022), Vivian Sobchack (2006), Sarah Jain (1999) and others have argued, this ignores the bloody and bruising difficulty of the join of tech to flesh, sliding over the incommensurability Fazi defines (a move that may be desirable in industrial design, which seeks to make things feel easy or easeful, but not in theory, which is seeking to understand, or account for this incommensurability). Instead, this paper explores various forms of contagion: the apparent leakage that shouldn’t happen, the form that shouldn’t appear vital but does, the uncannily close and also distant, the two-way connections that are made – somehow – but aren’t always about metal meeting flesh and may arise in the imagination, or via the persistence of vision, through mediation, or even as a spark jumps – not randomly, that is to say, and not in ways easily mapped, but with effect. Honesty here relates to how these connections present themselves perhaps, and what they do for real.
Intimations of this aesthetic already circulate widely. Consider the myriad AI-generated images that are ‘almost perfect’ simulations of the human but for their characteristically fused fingers, common enough to have become a shorthand for AI glitch in general (Wexford, 2023). Then there is the tail of the horse the impossible AI astronaut rides, hard and dead , like a a stone.5 The forms and practices of the new digital afterlife discussed here, which perhaps constitute case studies, partake of this aesthetic or even advance it, pushing hard at the tension between simulation and convergence (when does the circuit close, how does the spark leap or the parallelism bend, and finally how and in what way may the dead come to animate the unliving?). This issue lies close to the surface in Mori’s early account, although it has often been forgotten (set aside) in more general invocations of the artificial uncanny. Consider the diverse figures that ‘people’ Mori’s valley: a puppet, a zombie, an artificial hand (prosthetic and myoelectric, but synecdochic either way), a mask, a doll, a stuffed animal, a healthy human person, a corpse. The proliferation of forms, in their disparate materiality, is startling, as is the variety of the states they are in: dead, alive, unliving, immortal, whole, in parts. The ‘healthy human’ sits atop the valley, while the corpse (closest to the human, furthest from standard ‘life’) is deep in the ravine where the mixing threatens to happen.
It is clear that what is uncanny here is not only machines coming close to life, but humans becoming contaminated by, or coming close to, or mixing with what might be contagiously proximate forms of half-life, unlife, dead life. What is disturbing, that is, about the valley encounter is not only that life meets near-life, but also – and this should be no surprise if we have read Sigmund Freud (2003; see also Jentsch, 1997) on the uncanny – that life is coming too near to death to be sure of itself (Mersch, 2023; Giddings, 2024). Glossing Mori’s description of the moment of revulsion, Dieter Mersch (2023: 156, my italics) homes in on the corpse, noting ‘that which is too similar becomes dissimilar and revolting, like a corpse that is no longer the same as a living body … sparks a mixture of revulsion and fascination’. At the base of Mori’s valley are things that together, half-alive or half-dead, unsettle and render contingent naturalised boundaries that are conventionally supposed to divide life from unlife. It isn’t deadness or aliveness, then, but an accelerating ‘undecidability’ that produces the uncanny, one that is moreover accelerated further when objects are set in motion. Mersch argues that this is best defined specifically as an ‘aesthetic of decidability/undecidability’, and on this basis he suggests that the valley might constitute a kind of Turing test for the ears and eyes. My sense is that this final turn (towards Turing) won’t do, at least if it leans towards simulation and towards the problematic tendency, evident in at least latter-day appropriations of Turing’s test, to apply human measurement criteria to (judge) non-human activities, qualities and possibilities, including possibilities for inauguration, the production of the new. Here I am back to discontent with a particular kind of reversion to the anthropomorphic which responds to its other, a discontent with the metaphysical, but in so doing loses sight of the technical. I argue, then, that what the uncanny valley points to, or can help to identify, is an emerging aesthetics of entanglement – which is not reducible to simulation – between that which is living and that which is not, that which calculates and that which feels. Informing this is both the irreconcilability between the organic and the calculable already briefly introduced (it is the artificial that produces the revulsion) and the fact that despite this, these bodies, the new inhabitants of the depths, nonetheless rub shoulders with each other – exchange vital matter, or donate it, in ways that might change both sides.
Mori (2012), it is to be remembered, suggested staying away: roboticists, he said, should ensure a hygienic distance from the contagious zone by modifying their design principles and practices. Nobody listened, and the proximity he warned against is now routine in virtual spaces, if not yet observed in relation to robots.
It is particularly marked in relation to the digital afterlife and digital afterlives, which constitute the matter of the remainder of this paper. Consider the proliferation of new figures of the digital afterlife: fictional clones, deadbots, digital twins, griefbots, animated ancestors, revenant celebrity authors, dead constructs. All of them fall into the loose category of thanobots. They don’t stay quiet like the dead are supposed to do, although they’re not quick either. At times it seems their motion is designed to reflect this, as when the image of a man long dead offers, or seems to offer, the slow rictus of a smile. Thanobots have been usefully taxonomised in the register of speculative design by Thomaz Hollanek and Katarzyna Nowaczyk‑Basińska (2024); they have also become the subject of multiple ethical considerations (see Hollanek et al., 2024; Lim, 2024), come under increasing scrutiny in various ethnographic accounts, and are linked to long-standing death studies. They are also the subject of much popular discussion, the latter characterised by a queasy fascination, which, in its familiarity, points to that which has remained unresolved in popular culture about the issues Mori raised. Thanobots intrigue because they personify (possibly) a rearising undecidability about what constitutes or animates life in AI times, about the status of things that appear half-alive, and about what might be the terms and materials, even the stages, of an exchange through which one form might animate another.
A strand of conspiratorial netlore raises the spectre of the dead internet, full of bots but empty of ‘us’, the murderous event supposedly having taken place in 2016 or thereabouts. In my view what needs to be contemplated instead – what both constitutes or materialises the new aesthetic already under discussion and provokes consideration of how or what shape it might take – is the undead internet.
Differentially positioned, but understanding ourselves as human, we might listen for the clamour of active dead things. Or we may watch ourselves enacted at a distance, warranted by a remote feed to which we contribute for a time, or unwarranted, sometimes in violent ways, as collective activity overlays something that was once our own This is a terrain where the forms and practices, the activities of life and death, of the dead and the living, or at least the unliving or the artificial, are exchanged in various ways and with increasing frequency, all this being at once deeply uncanny (its ‘creepiness’ being widely acknowledged in popular culture) but also banal – a characteristic bathos constitutes the standard mode of address of the afterlife industry. The thanobot imaginary might draw on legends of immortality, on a vampiric (undead) or technological sublime, but it is also enacted, grounded, practical, familial rather than mythical. Since computational capitalism has become compulsory (Bassett, 2022; Woods, cited in Bassett, 2022: 57), we will increasingly live in the valley it has made, even if we wish we did not. Furthermore, the afterlife industry will make it their business to domesticate things for us (see e.g. Montag, 2024) Or rather, for them. The bottom line here is expansion: lives, histories, memories, pasts, futures, capabilities, capacities to attend, are all stretched backwards and forwards, further than a standard lifespan can go. If the internet was (claimed as) the old digital frontier, now there is a new one. This one is not based on (the fantasy of) expanding into limitless space, but on expanding the span of life, so that what inhabits that space might remain a data point for longer: the empire of death is, what was previously after life is, through computation, to be settled. This expansion shouldn’t be a surprise; it only constitutes another turn of the same screwand claims it is unlimited are exposed as a fantasy by the same measure through which the fantasy of the limitless grounds of cyberspace was exploded: it only requires consideration of the environmental cost of the supporting infrastructure.
In invoking ‘us’ here I am rather consciously tailing popular discussion about deadbots, thanobots and automation anxiety, but that mode of address should be interrogated. The assumption that there is something universal, or some universally shared point, that indicates the moment when technology approaches life, when simulation is near perfection, or life is near unlife, or even when unease ought to occur, needs to be questioned. The skin is faster than the word, declared Massumi (1995: 25), and in doing so he opened up new ways to think about the affective dimensions of culture, but the presumption that these responses are instinctively produced, or universally valid – ‘human’ – needs to be pressured: the ‘factor’ in a yuck factor is often socially constructed and defined as natural by those with the most construction power (Bassett, 2006). From what mode or image of bodily ‘perfection’ is it that ‘we’, or the afterlife figures ‘we’ shudder at, deviate from? In her work on thanobot ethics Ella Lim (2024) argues for a consideration of human dignity. I would argue, as part of that, or as an extension to it, for a recognition of human difference.
Late in his life, reflecting on his work, Mori came to reject the placement of the corpse in the depths of the valley and the ‘healthy’ human at the top, arguing in favour of forms of post-humanism that refused that hierarchy, and also rejecting the locus of the body itself as the secure centre of things (Kageki, 2025). Mori’s revisionism avowedly related to the development of his Buddhist beliefs, but it points to other ways to think about the body – or to question assumptions about the binary division between bodies and machines, and about bodily boundaries, now unsettled, as having been previously secure. The dream of the perfection of the human in a new era of the technological sublime, configured by some as the perfection of simulation, as a moment that paradoxically entails complete convergence and total division (the human becomes machine via uploading, or the machinic extension of the human becomes seamless via idealised prosthesis), is undermined if entanglement is already assumed, and if – while life may be desired and death feared – their relation is differently conceived of from the start.
This again indicates why working in the register of simulation (looking at simulated selves, lives, memories, bodies, identities, personalities) alone doesn’t capture the complexity of these emerging formations; above all, it can’t help us investigate that their force emerges from how they connect.
Pursuing these issues, this paper, which began with Mori’s speculative diagram, now briefly alights on a fictional spaceship (in Olga Ravn’s The Employees, a novel about a future workspace), considers the undead and the changing tense of photography, and finally looks at digital twins and the offer of inert immortality to those who work hard enough on their own (re)production.
Ship Life
Ravn’s (2018) The Employees: A Workplace Novel of the 22nd Century crystallises in fiction the peculiar admixture of the banal and the uncanny that is the operating aesthetic of the digital afterlife in the real. In her quasi-documentary-style novel, humans and artificial clones, confounded by each other’s sameness and difference, share a pull – an ‘attraction revulsion’ – towards mysterious, entirely alien objects, whose influence expands as the employees continue their work routines (canteens, checks, small hierarchies). Something catching, some uncertainty that layers onto what is already an ontological unease about selves and others (about human-shaped life and human-shaped artificial life), spreads across the doomed employees. As employee 052 puts it: ‘I feel a … longing to be a human, as if somehow I used to be, but then lost the ability’ (Ravn, 2018: 45).
Discussing the biomachinic body, taken as the body we now inhabit, Eugene Thacker (2003: 53) declares that it ‘proceeds via a dual investment in biological materiality, as well as the informatic capacity to enhance biological materiality’. In this way the ‘viewpoint of bioinformatics gives us a radically different notion of biological normativity and health’ (Thacker, 2003: 77) and produces an account that is completed in the biological (post)human, even if there is no pristine flesh, and no pristine way to understand ‘life’ that is not already technological, to be found.
These days, given deadbots, griefbots and digital twins (and perhaps the rise and fall of 23andMe), perhaps this certainty of an end point (‘completion’) is eroding. Ravn’s employees wonder about this. If humans are biomachines, clones are also half-human and half-technology, so how can the division between them, which is felt, be made, and in what does it inhere? Is this a matter, like being the right size, of having the right ratio? ‘Half human. Flesh and technologyToo living’ (Ravn, 2018: 16) comments a clone, and it is a question as well as a statement, perhaps also a complaint, or even a lament. Clones and humans, knowing their designated state, appear increasingly undecided about what their designation means, about what separates them from each other. This is an uncertainty that is passed on to the reader, who does not always know either – and is not intended to: ‘So I will make an Android and a human, and I will make them both speak but I won’t tell the reader who is speaking’ (Ravn, cited in Ellison-Balaam, 2021). Even immortality (or future re-decanting) seems to make very little difference to the clones, who will all experience an end to their employment, which constitutes a kind of lifespan. It isn’t what they regard as decisive in their difference from humans, nor even what gives them their greatest (perceived) advantage. Little in The Employees is resolved. Evidence statements only provide glimpses of the termination of the programme, although one thing is clear: the objects – ‘forms that were not really human, but still living’, as Ravn acknowledges (Ellison-Balaam, 2021) – whose status is entirely undecided, and who fall out of any human/clone binary, confound any categorising of the organic biomachine versus the clone, even the fleshed clone, and contribute to the confused sensations all report.
Why turn to Ravn? Or to fiction? After all, it might be thought there is enough unlife in real ‘life’ these days, including the real afterlife industry and formation under discussion. My reason is because the forms of contagion I am exploring include or provoke questions about the relationship between fictional and real lives – and deaths. Sandy Stone (1995), the 1990s theorist of the internet, desire, technology, bodies and lives, commented that she found the shift between fiction and the real to be ‘a kind of uneasy thing’ but also one that she found productive to investigate, partly as a theoretical approach, but also perhaps because that shift, or living the shift, already constituted an integral part of the then emerging first wave of the internet as a techno-cultural formation – and is still notable today. In The Employees this fiction/real shift is articulated through a contradiction in form (notably via the documentary reports from the numbered workers, which appear ‘factual’) and also emerges as content – for instance, in the anxiety around the fabulated and the unfleshed, or in how humans are and/in relation to the artificial life that shares a canteen with them and occasionally kills them.
Thinking of the work and the world, I would argue that The Employees gathers up and articulates something of the political unconscious of computational capitalism at a specific stage, constituting a particular kind of depth investigation of what might be termed artificial stupidity (O’Connell, 2017), which is shown to emerge not as something characterising either human or machine workers (no IQ comparisons) but as a feature of routinised, largely meaningless labour, the purpose of which is occluded. However, there is another way that fiction can be useful to this enquiry: it offers a means through which an incommensurable experience might be grappled with, and one that is useful here. Enquiring into literature’s potential to investigate materiality (or to work from a new materialist perspective), Tobias Skivern (2018) considers Anne Alaimo’s claim that fiction offers a means through which that which may not be seen may be made palpable, enabling what he terms ‘transcorporeal flows’ to be comprehended even if – or in spite of claims that – literature might seem poor grounds for exploring forms of (vital) materiality.
Ravn’s workplace troubles the alignment of human bodies with life and of machines with the unliving, gives us biomachines of distinctive kinds, and also points to the impacts of their contagious intersecting, with each other and with non-humanoid agencies – the vibrant things on the ship, in fact inspired by real-life sculpture by Lea Guldditte Hestelund (Ellison-Balaam, 2021). It thus enables us to speculatively palpate some of the questions at issue here, to grapple with an aesthetic that is at once pervasive (even constituting an atmosphere), iterative (repetitive, ongoing, cyclical, numerical) and characterised by severance and reconnection: a form of contagion that leaps or bleeds, but not symmetrically, between different bodies and in multiple situations. Can a clone be a biomachine? (Or, as we will come on to, can a digital twin amount to that – after the death of its donor, but having had a body?) How much flesh, or life, or distance, or closeness does it need? What ratio must persist, or not?
Deep Nostalgia, or a ‘Variation in Movement’
When the speed is cut in half in an attempt to make the robot bring up a smile more slowly, instead of looking happy, its expression turns creepy. This shows how, because of a variation in movement, something that has come to appear very close to human – like a robot, puppet, or prosthetic hand – could easily tumble down into the uncanny valley. (Mori, 2012: 4)
Deep Nostalgia, offered by genealogy service MyHeritage, deploys an AI algorithm to animate old photographs. It will work on your ancestors for you: brush one up against the other, the algorithm against the image, and life is forced out of the mouths of the dead – or at least their lips move. The sendings of the nostalgia machine come near the one who sends for them, but they also remain far from life, aloof; the images promise revelation or communication but don’t provide it. There are no hints or advice from elders available; what’s on offer is only a kind of revivification; ‘an insanely lifelike state’ as one Reddit substack on genealogy put it.
A image of my great uncle surfaced on the internet decades ago, via British Army archives perhaps, or perhaps it was uploaded as part of a genealogical search. However, it has been familiar to me, as a framed photograph, all my life. My uncle was called Noel. He was 22 when he died at the Somme during the First World War, shortly after this likeness was taken. Many soldiers did the same before they went to the front, where they knew they would almost certainly die, and this death is anticipated in the photograph. Noel is beautiful, a wax figure with a grave face.. Roland Barthes’s (2010) famous argument that the photograph’s tense is ‘that has been’ seems peculiarly apt in this case. I consider submitting this image to Deep Nostalgia, the offer being to ‘bring your cherished memories to life’. I don’t. Rather, I submit to the algorithm an unknown ancestor, a cricketer who now, digitally revenant, offers me a slow sardonic smile. At once still dead, never living and faintly alive, this image has no breath of life, no lungs behind the smile. Barthes (cited in Grundberg, 1981), about to die himself, and reflecting in Camera Lucida on the death of his mother and on photography and memory, distinguished between stadium and punctum: the punctum is that which breaks through the ordinary – pricks me/bruises me/is poignant to me – and it might constitute an ‘emanation’ of a kind, although Barthes was clear that photography was chemistry, not magic. Grundberg (1981), in an early review of Camera Lucida, called the whole book a meditation linking photography to death. If so, it also links to life, or to the life that has been lived. Benjamin (1972: 7), whose ‘Short History of Photography’ is cued by Camera Lucida, discusses what the photograph holds onto and won’t give to art – and perhaps not to anything else either – doing so through the fishwife of Newhaven, a woman who will not ‘yield herself up entirely’.
While placing the life in the image as that ‘lived at the time’, Benjamin (1972: 7) also argues that this life, or this fishwife, remains ‘real even now’. Is the image then vital or dead? Benjamin talks of the ‘tiny spark of chance’ looked for in the photograph by the viewer, that which might be enabled by the close-up/aura-busting eye of the camera: a spark of the here and now, something re-discoverable, unfinished, that also constitutes a future to come. It may be this that also leads him to suggest how this links to desire for possession, noting that ‘every day the need grows more urgent to possess an object in the closest proximity’ (Benjamin, 1972: 20).
Deep Nostalgia might be said to constitute a refusal of death in the image, but if it does, it might also be said to negate the chance of discerning in it the Benjaminian spark. Does the animated image lose its spark, the intensity with which it was there – how its presence was – in favour of some other form of liveliness? This seems worth considering. Deep Nostalgia functions not through the kind of prompt that evokes a memory of a person, but more directly, or in a new way, by bringing the object (the human subject) of a memory back into presence. It might thereby be fuelled by (my, your) deep nostalgia while also commoditising the offer of closer possession. Perhaps what is gruesome, or causes a moment of revulsion, a hesitancy here, is not death at all, or reanimation, but the violence underpinning the drive to colonise a memory. There are puppets in Mori’s valley. My reluctance or unease about creating them, at least where they relate to those I know to have lived, arises in part because the process seems dishonest, and also faintly coercive, as others have noted in similar cases, both fictional (see for instance William Gibson’s (1984) Neuromancer) and in the digital real (Sherry Turkle’s (2011) explorations of puppetry are of note here).
Dead Ringers
They would have left behind a part of their essence that people could access at any time through AI-based platforms.(Oliver, 2-25)
From the dead other to the dead self. Digital twinning is commonplace and has multiple uses. On prediction sites or in test beds, objects or humans (or processes) may indiscriminately be the subjects of twinning, but there is something distinctive about twinning as an afterlife procedure or service. The afterlife twin is explicitly offered as more than an informed search engine (based on a database or archive of a life), more than a static modelling package (say, one that models a personality), and it claims to be, or appears to be, more than a simple simulation (iterating actions based only on its dedicated training/training resources).
The twin can also be distinguished from earlier internet forms and practices to which it obviously relates, being the logico-cultural extension of social media platform presence, life-logging, blogging, and the quantified self movement. The autofictional turn responds to these developments, perhaps; I would argue that Ravn’s The Employees can be understood as a mode of fictional autofiction.
What the twin shares with these cultures is perhaps the claim to be contributing going forwards. The twin is intended to simulate (probabilistically) a person’s future responses to a set of questions or queries, and in this way it is distinguished from, is supplementary to, a static archive of that person’s thoughts, not least because it continues to build, to operate, to traffic. The offer is thus of a form of spectral agency, a form of identifiable personhood that can operate after the person’s death (and part of the allure might be the hope that the twin can control the person’s reputation, perhaps, or temper how a person is known – continues to be known – in public after the death of the body). But does the twinbot have enough liveliness to produce the new? In computational terms it is enjoined to do so (it must not simply repeat or overfit), but it also must not (it must be recognisably true to its original and/or good enough to seem or to be used in that way). What would be the point of an ancestor who changed their mind, or the point of leaving a monument to one’s bigotry if the monument got woke? In fact, the offer tends to remain personal or intimate: these services invite you to ‘turn a loved one’ or even a ‘dying relative’ into an interactive twin, or to make a twin of yourself, explicitly sometimes to intervene in grieving processes.
The instructions on how to twin vary: ‘microdose your double at regular intervals’, advises one site selling a twin service (Griffin, 2024). The vampirical resonances of this might imply a diminution of organic life as it feeds its unalive double (the stuff of fiction and myth), but the real offer does not depend on an exchange of vitality that weakens one side, but on image control and presence into the future, a hedge against posterity and its revisions. Through this form of ghostly presence the twin offers to provide a means to continue to contribute, or even to continue to exercise control: ‘The creation of a digital twin allows a person … to maintain interactions with family and friends posthumously’. These are the kinds of ambiguous claim – I mean this both in the sense that they are unsustainable and also that they might imply coercion – widely made by various hawkers of long life, or rather of a long and active death.
Perhaps an active death is a half-life. Either way, it might continue indefinitely, since the peculiar promise of afterlife twinning is to deliver a form of post-mortem immortality that is simultaneously acknowledged to be an illusion. ‘Through digital twin technology people could ask questions and have conversations which would create the illusion that their beloved person remains alive digitally’ – so says an article with the headline ‘How AI Could Let People Converse with Departed Loved Ones’ (Oliver, 2025). The kind of self-produced immortality this enables, not delivered via a cryotank or a Moravec-style upload situation, amounts in some ways to the bathetic re-inscription of the hubristic desires underpinning those (fantasised as realistic) routes. One difference is that twinning works – at least if, or insofar as, what is expected is the generation of a ghostly presence that is apparently coherent, with a store of memories as data to draw on, and which is able, given this, to deliver, in probabilistically derived ways, interventions consonant with a person(ality). Hollanek and Nowaczyk‑Basińska (2024) develop intriguing scenarios exploring what happens when the personality and its life don’t match those its later interlocutors recognise or expect.
What makes afterlife twinning distinctive is the specious (and inconsistent) but oddly powerful claim – powerful because backed up through various forms of warranting, and this does not reduce to prosthesis – that the twin might, in some way, be ‘you after you’, a part of the ‘essence’ of you being saved.If afterlife services claim to have something to do with or to have caught something of life (claim a link to it, even if that link is to the uncannily living dead), that claim is supported by the sense that the twin can contain or articulate a personality or constitute a form of revenant personhood.
This invites a consideration of what that personhood inheres in, and here the distinction between ipse and idem as aspects of personhood guaranteed by continuity across time and/or by faithfulness to the self – another articulation of the continuous and the discrete, but one that also points to something like the disjunctively continuous – can usefully be invoked (Ricoeur, 1990; Bassett, 2014; Arendt, 1998). These distinctions begin to suggest another way to think about simulation (the making of the same) and the possibilities of the new or inaugural within the grounds of something recognised as also claiming a different kind of consistency, one that, edging away from the metaphysical, moves towards something more expansive or grounded, lets back in questions of exploitation and ethical engagement (perhaps even beginning to gesture towards a media ethics of distance and closeness).
The twin, remember, is intended to simulate (probabilistically) a person’s future responses.It is to be distinguished from, or is supplementary to, an archive of that person’s thoughts, because it continues to build, to operate, to contribute even, to traffic in the world or even be trafficked. In this sense it might be not only dead but alive, and it might claim some kind of personhood. Such questions about the status of the sub-self are already palpated in fictional form: this is Severance’s plot; at one remove, it is the plot of Kazuo Ishiguro’s Never Let Me Go; it is the question of interest in a legion of clone sci-fi tales dealing in the afterlife of backed-up selves.
It might be that when the disavowed security that enables the claim of life to be understood as only metaphorical is disturbed by the apparently lively and personful operations of the agent, the deeply uncanny arises, provoking the moment of revulsion. This is when the subject who ‘had been’ there (Barthes, 2010) but is now returned looks as if it has been cajoled into becoming the object present at hand as a standing reserve (Heidegger, 1977), reduced to being like any other object.
The twins are the algorithmic other of the living body and so might constitute the perfect test bed for claims regarding an essential distinction between the aesthetics of a machine life and the aesthetics of an (already) biomachinic body. But in operation what is evidenced are ambiguous, disjunctive, multilayered forms and sites of connection, contradiction, contagion and exchange, and these seem central in deriving an aesthetic of emerging biomachinic life (and death) that does not cleave to one side or the other of the living or dead, the human or machinic, but lives – uneasily – in the uncanny between. That is, it may be possible and useful to dwell on the apparently incommensurable schism between the discrete and the continuous, and also to point to the consequences of cleaving to one side. I am referring here to Fazi’s (2019) argument. This would suggest that (human) life, which guarantees one form of aesthetics (bodily sensation), is routinely (maybe even unjustly) prioritised, so that the potential for exploring what might be the formal aesthetics of computation is lost. But through the examples explored above, I have also pointed to processes through which mediation (let’s say) nonetheless takes place.
If this is possible, it shouldn’t be a surprise. Shifting from contagion to mediation, or regarding the former as undertaking the latter, is useful here since it points to the theoretical resources of medium theory. This has taught us that there are all kinds of ways in which sensation cheats ontology – and specifically by way of forms of animation. What is cinema, if not, as Kittler (1999: 120) or Fechner (cited in Kittler, 1999: 120) told us, a way to render, through the persistence of vision, the discontinuous discrete, something demanding of technology but also relying on the eyes of the human – a contagious process that also demands movement to come alive? In the case of the living dead, it is memory – a technology no less than the persistence of vision that enabled film – but also our own capacity to understand personhood in ways that tolerate the discontinuous as well as the discrete, or that can switch between different ways in which that is measured or expressed, that can make the bridge. That is why what needs to be further explored in the ditch, in the valley bottom where we are now, and where we are treated in uneven ways, is not to be reduced to a matter of simulation. Rather, this confluence of vitalism, animation and operation might constitute an aesthetic that has to be both human and machinic. Contagion in the end is a distinctive form of mediation; flesh and technology in newly uncertain combination, and digital afterlives, might suggest the rise of new forms of the biomachinic, which while threatening to increase control, and while expanding the realm of expropriation, might also have potential. There is something more here than a Musky route to a dead planet when we have killed our own.
References
Arendt, H. (1998) The Human Condition. Chicago: University of Chicago Press.
Barthes, R. (2010) Camera Lucida: Reflections on Photography. London: Hill and Wang.
Bassett, C. (2006) ‘Yuck Factor’, in C. Jones (ed.), Sensorium. London: MIT Press.
Bassett, C. (2014) The Arc and the Machine, 2nd ed. Manchester: MUP.
Bassett, C. (2022) Anti-Computing. Manchester: MUP.
Bassett, C. (2024) ‘The Author, Poor Bastard: Writing, Creativity, AI’, in W. Slocombe & G. Lively (eds), The Routledge Handbook of AI and Literature. London: Routledge.
Bassett, C (2025) “Cyberfeminism”, in the Bloomsbury Encyclopedia of New Media Art. London: Bloomsbury.
Bassett, C., & Roberts, B. (2019) ‘Automation Now and Then: Automation Fevers, Anxieties and Utopias’, New Formations 98: 00–00.
Benjamin, W. (1972) ‘A Short History of Photography’, Screen 13, 1: 5–26.
Ellison-Balaam, R. (2021) ‘Reading with the Mouth: A Conversation with Olga Ravn’, Lolli (January 26). https://www.lollieditions.com/lolli-in-conversation/reading-with-the-mouth
Fazi, B. (2019) ‘Digital Aesthetics: The Discrete and the Continuous’, Theory, Culture & Society 36, 1: 3–26.
Freud, S. (2003), The Uncanny. London: Penguin.
Gibson, W. (1984) Neuromancer. London: Harper Voyager.
Giddings, S. (2024) Toy Theory: Technology and Imagination in Play. London: MIT Press.
Gillis, A. S. (2024) ‘What Is Uncanny Valley?’, TechTarget (February 6). https://www.techtarget.com/whatis/definition/uncanny-valley
Griffin, M. (2024) Grief Tech Uses AI to Help You and Your Loved Ones Become Digitally Immortal, 311Institute. (https://www.311institute.com/grief-tech-uses-ai-to-help-you-and-your-loved-ones-become-digitally-immortal/)
Grundberg, A. (1981) ‘Death in the Photograph’, New York Times (August 23).
Haraway, D. (1991) Simians, Cyborgs, and Women: The Reinvention of Nature. New York: Routledge.
Heidegger, M. (1977) ‘The Question Concerning Technology’, in M. Heidegger, The Question Concerning Technology and Other Essays. London: Garland.
Hollanek, T., & Nowaczyk‑Basińska, K. (2024) ‘Griefbots, Deadbots, Postmortem Avatars: On Responsible Applications of Generative AI in the Digital Afterlife Industry’, Philosophy & Technology 37: 63.
Ishiguro, K. (2005). Never Let Me Go. London: Faber and Faber.
Jain, S. (1999) ‘The Prosthetic Imagination: Enabling and Disabling the Prosthesis Trope’, Science, Technology, & Human Values 24, 1: 31–54.
Jentsch, E. A. (1997) ‘On the Psychology of the Uncanny’, Angelaki 2, 1: 7–16.
Kageki, N. (2025) ‘An Uncanny Mind: Masahiro Mori on the Uncanny Valley and Beyond: An Interview with the Japanese Professor Who Came Up with the Uncanny Valley of Robotics’, IEEE (January 24). https://spectrum.ieee.org/an-uncanny-mind-masahiro-mori-on-the-uncanny-valley
Kittler, F. (1999) Gramophone, Film, Typewriter. London: Stanford University Press.
Lim, E. (2024) ‘From Mourning to Machine: Griefbots, Human Dignity and AI Regulation’, Commentary Solutions (August).
Litchfield, T. (2024). ‘There’s Already a Company That Will Turn Your Dying Relatives into Perverse AI Homunculi for $50K’, PC Gamer (May 17). https://www.pcgamer.com/gaming-industry/theres-already-a-company-that-will-turn-your-dying-relatives-into-perverse-ai-homunculi-for-dollar50k/
Massumi, B. (1995) Parables for the Virtual. London: Duke University Press.
Mersch, D. (2023) ‘Get into the Uncanny Valley’, in D. Mersch et al. (eds), Actor & Avatar: A Scientific and Artistic Catalog. Bielefeld: transcript. 156-164. https://doi.org/10.1515/9783839467619-021
Montag, M. (2024) ‘This Company Will Turn Dying Relatives, Loved Ones into Interactive AI “Twins”’, Penn Live (May 21). https://www.pennlive.com
Mori, M. (2012) ‘The Uncanny Valley: The Original Essay by Masahiro Mori’, IEEE Robotics & Automation Magazine (June).
O’Connell, M. (2017) Art as ‘Artificial Stupidity’. PhD thesis, University of Sussex. https://hdl.handle.net/10779/uos.23443571.v1
Oliver (2025). AI and Digital Twins Could Let us Talk to Departed Loved on, Says Google X founder. TechsterHub. (Techsterhub.com).
Ravn, O. (2018) The Employees: A Workplace Novel of the 22nd Century. Copenhagen: Lolli Editions.
Reichardt, J. (1978) Robots: Fact, Fiction and Prediction, London: Penguin
Reichardt, J. (1968) ‘“Cybernetic Serendipity” – Getting Rid of Preconceptions’, Studio International (November 1). https://www.studiointernational.com/index.php/cybernetic-serendipity-getting-rid-of-preconceptions-jasia-reichardt
Ricoeur, P. (1990) Time and Narrative, vol. 3. London: University of Chicago Press.
Erickson, D. (2022) Severance, AppleTV.
Skivern, T. (2018) ‘Literature’, New Materialism (February 7). https://newmaterialism.eu/almanac/l/literature.html
Sobchack, V. (2006) ‘A Leg to Stand on: Prosthetics, Metaphor, and Materiality’, in J. Morra and M. Smith (eds) The Prosthetic Impulse: From a Posthuman Present to a Biocultural Future. Cambridge: MIT Press.
Sterne, J (2022). Diminished Faculties. Durham: Duke University Press.
Stone, S. (1995) The War of Desire and Technology at the Close of the Mechanical Age. London: MIT Press.
Thacker, E. (2003) ‘What Is Biomedia’, Configurations 11, 1: 47–79.
Turkle, S. (2011). Alone Together: Why We Expect More From Technology and Less From Each Other. New York: Basic Books.
Wexford, C. (2023) ‘Why Do AI People Have Extra Fingers?’ Medium (December).
Endnotes
1. DeepBrain AI claim widely circulating across specialist streams (Montag, 2024; see also Litchfield, 2024).
2. Both the exhibition and the lectures that dealt with the main theme of ‘cybernetic serendipity’ were designed to dispel some of the prejudices that cause many intelligent people to believe that the computer is a threat to those whose intellectual abilities and creative powers may no longer be in demand as their role is gradually taken over by AI machines (Reichardt, 1968).
3. Seduced and Abandoned: The Body in the Virtual World.
4. ‘Our machines are disturbingly lively, and we ourselves frighteningly inert’. Haraway’s (1991: 149) warning, made some years later, on the brink of the virtual age, was a call for a critical response to cyberculture, but it was couched in quasi-ontological terms.
5. The horse on the moon was a stable diffusion prompt that circulated widely.
6. To learn more about how to build a chatbot that can avoid the uncanny valley, see Gillis (2024).