The Limits of Digital Immortality
M/C Journal · 18.06.2026 · 29 (3)
DOI: 10.5204/mcj.3285 ↗Kurz gefasst
Fachartikel in der Zeitschrift M/C Journal, online erschienen am 18.06.2026. Verfasst von Eva Nieto McAvoy und 2 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0).
Abstract (Originalsprache)
“We Never Have to Say Goodbye”, promises You, Only Virtual, a commercial platform whose service claims to reconstruct a person from their digital data and make them available for interaction after death. You, Only Virtual is one of a growing number of commercial services—alongside Almaya, Here after AI, or Seance AI—that have come to be grouped under the informal designation of “deathbots”. Deathbots are AI-mediated interactive simulations of the dead, constructed from aggregated personal data (Bao and Zeng), and oriented toward what their proponents frequently describe as digital immortality, a term that carries considerable cultural and philosophical weight. As Stephen Cave has argued, the desire to persist beyond biological death organises some of the most enduring narratives in human culture, from the resurrection myths of antiquity to the soul theories of the Enlightenment. Deathbots are the latest iteration within this long history of the symbolic immortality of memory, legacy, or creative work that are the dominant modes through which humans have historically managed the fact of death (Lifton). But they also promise something ontologically different. Deathbots are framed as allowing the self to continue after death as an interactive digital presence. It is precisely this claim that this article sets out to examine, mobilising deathbots as a critical case study through which to interrogate what kind of immortality digital systems can produce and with what consequences for how we understand the persistence of the self after death. We do not argue simply that deathbots fail to deliver on their promise, as something clearly does persist in these encounters, data do outlive the person, and users do experience something that warrants serious critical attention. We argue, rather, that digital immortality can be understood in this context as the dispersal of the self into fragile and platform-dependent traces, where our data persist at massive scale (Öhman and Floridi; Öhman and Watson; Öhman), while the continuity of any individual person remains partial, fragile, and unevenly distributed. Deathbots vary considerably in their technical architecture and commercial form, from services that generate text-based conversational agents trained on messaging archives and social media data to more sophisticated multimodal systems that incorporate voice, image, and video to produce interactive avatars (Kidd and Nieto McAvoy, "Synthetic Afterlives"). What they share is a dependency on data captured during the lifetime of the person they simulate, which is aggregated and used to train models capable of generating responses that approximate the communicative patterns of the deceased. Increasingly, this data capture is not only post-mortem but anticipatory as some platforms actively encourage users to contribute data before death. This enables a degree of curation and editorial control through which the future dead person can, in effect, shape and edit the version of themselves that others will encounter after they are gone (Kidd and Nieto McAvoy, "Synthetic Afterlives"). Such pre-mortem construction introduces a layer of intentionality and agency that distinguishes some deathbot practices from the passive accumulation of digital traces, though it does not resolve and, in some respects, deepens the questions this article addresses, particularly the nature of the digital self. Moreover, the infrastructural conditions underlying deathbots are marked by a structural asymmetry that complicates any straightforward account of digital immortality. Platforms accumulate these fragments of the dead in quantities that will, on current trajectories, come to dwarf those of the living (Öhman and Watson). However, while data traces persist at scale (Öhman), the continuity of any given individual, we argue, remains fragmentary and precarious, not least of all because they are dependent on ongoing technical and economic support that is neither guaranteed nor permanent (Knees; Öhman). The benefits of this data persistence are also not democratically distributed, as the labour and costs fall on the dead (and their living relatives), while most of the benefits (at the very least, financial) accrue to the platforms that hold and monetise digital remains (Öhman and Floridi; Öhman). Scholarship on deathbots and adjacent digital afterlife phenomena has begun to emerge across questions of personhood and dignity (Stokes), continuing bonds and the emotional implications of these practices (Fabry and Alfano; Brescó de Luna and Jiménez-Alonso), the design and governance of AI afterlife platforms (Morris and Brubaker; Hollanek and Nowaczyk-Basińska; Kneese), and the mnemonic and affective dimensions of synthetic afterlives (Kidd and Nieto McAvoy, "Synthetic Afterlives"; Kidd and Nieto McAvoy, AI Afterlives). This article builds on and advances this scholarship by offering a theoretically informed and empirically grounded analysis of a selection of deathbot platforms. Our arguments are based on findings from a series of platform walkthroughs and discourse analysis of publicly available documents from four companies: Almaya, HereAfter, Seance AI, and You, Only Virtual. We argue that the infrastructural fragility of deathbots unsettles the claims to digital immorality offered by these platforms, and that what sustains this presence are the unstable and negotiated forms of engagement in which the encounter with the digital object is emotionally resonant but remains ontologically uncertain. At the interface level, users encounter what we have elsewhere termed contested presence: an engagement that remains ambivalent and liminal, co-produced at the intersection of the system’s outputs and the user’s memory of the deceased (Kidd and Nieto McAvoy, AI Afterlives). These encounters are ethically and philosophically unstable, not only because their meaning needs to be constantly negotiated, depending heavily on the emotional investment of users, but because they are shaped by inconsistent data traces and uneven platform infrastructure (Öhman, Kneese). In previous work, we have studied the affective and relational dimensions of deathbots (Kidd and Nieto McAvoy, "Synthetic Afterlives"). In this article, we focus on the infrastructural conditions that underlie and constrain whatever form of immortality these systems try to deliver. In the remainder of this article, we (1) situate deathbots within the longer cultural history of immortality, identifying what is distinctive about these systems, particularly in relation to the forms of digital personhood they articulate; and (2) examine the infrastructural conditions and the asymmetries those conditions produce; before (3) concluding by reframing what deathbots reveal about the nature of digital immortality. The Immortality of the Digital Self: Between Legacy and Resurrection The desire to persist beyond biological death is among the most enduring preoccupations of human culture, and the forms through which that desire has been articulated have shifted considerably across time and context, even as the underlying impulse remains remarkably consistent (Lifton). Cave’s taxonomy of immortality narratives offers a productive framework for situating the deathbot within this history, identifying four broad modes—staying alive, resurrection, the soul, and legacy—each proposing a different account of what survives, in what form, and under what conditions. It is within the last of these that digital memorialisation has most commonly been situated; the photograph, the recorded voice, and the archived message are understood as extensions of a long history of material practices through which the dead are made to continue in the lives of those who remain (Recuber; Kasket). Deathbots, however, make a claim that exceeds legacy and moves toward something closer to Cave’s resurrection narrative, promising to animate traces rather than merely preserve them, and to sustain something that responds in real time rather than remaining available for others to interpret. They occupy, in this sense, an uneasy and largely unexamined position between legacy and resurrection. The modes of continuity through which humans have historically managed death anxiety are all, in different ways, predicated on the acceptance that what persists is a ‘symbolic immortality’ (Sherlock; Recuber), a representation or legacy, rather than the person themselves. Even Lifton’s fifth mode, experiential transcendence—the felt, immediate sense of continuity most commonly associated with spiritual experience—understands that continuity as exceptional, transient, and not routinisable. The deathbot’s claim aligns, however imperfectly, with the transhumanist tradition whose horizon includes mind uploading (Kurzweil), whole-brain emulation, and the singularity, and against which deathbots present themselves as an available approximation that requires only the accumulation of sufficient personal data. This alignment is both technical and ideological: the commercial and infrastructural conditions of digital immortality are presented as the natural and inevitable fulfilment of a universal human desire, obscuring the specific cultural formation, rooted partly in Silicon Valley’s techno-utopian imaginary, from which they emerge (Morozov). What underwrites these claims to some sort of immortality is a set of assumptions about the digital self. On the one hand, the ‘informational self’ (Floridi) is sufficiently coherent and stable to serve as the basis for reanimation; and on the other hand, its capture and reassembly by a technical system is equivalent, or near equivalent, to the continuation of a living person (at least their mind or soul). These assumptions sit uneasily alongside what work on the digital self has consistently demonstrated: that the digital self is never unified or coherent but always already distributed across platforms, contexts, and interactions, constituted through the systems and practices of mediation rather than simply expressed through them (Papacharissi; Turkle; van Dijck; Cheney-Lippold). The digital self is an ongoing process of assembly and reassembly, shaped by active choices like posting and deleting, where any sense of coherence is contingent and thus continuously changing. Generative AI systems attempting digital resurrection try to synthesise a unified subject out of posthumous data that were already fractured to begin with, as well as intertwined with other ‘communicative traces’ (Kneese) and data bodies (Kidd and Nieto McAvoy, "Deep Nostalgia"). It is here that analogue and digital legacy traditions diverge in ways that are consequential for the various understandings of immortality. The photograph, the letter, and the recorded voice have always been understood culturally as fragments, their partiality acknowledged and their status as representations of the person rather than the person themselves broadly accepted. The digital self, by contrast, has been culturally constructed as something closer to a continuous and interactive presence (Kneese), and it is precisely this cultural construction that deathbots operationalise, reanimating traces under the cover of an assumption about digital continuity that was always more ideological than real. What the deathbot attempts to reanimate was never, in other words, the kind of thing that reanimation in any meaningful sense could recover; a point toward which Stokes’s scepticism about digital continuation and Derrida’s account of the trace converge, though from distinct intellectual traditions and with different implications (Of Grammatology). The WhatsApp messages, email archives, and videos from which deathbots are trained are traces in the Derridean sense, remnants that bear the mark of a presence that is no longer there, but that never was either. These traces can neither fully persist nor can they fully disappear, haunting the systems that claim to sustain them (Derrida, Specters of Marx; Fisher). The Infrastructure of Digital Immortality The idea that deathbots can deliver a meaningful form of digital immortality rests, as the preceding section has argued, on assumptions about the coherence and recoverability of the digital self; but it rests equally on assumptions about the infrastructural conditions under which digital traces are held and made available that are no less consequential and considerably less examined. Infrastructure, as Starosielski has argued, is characteristically invisible until it breaks down; its conditions of operation are obscured by the very smoothness of its functioning. It is precisely this invisibility that allows deathbot platforms to present the persistence of data as natural and indefinite. The material conditions of that persistence are, however, marked by fragility and contingency as “platforms wither and die, just like people” (Kneese). They are marked as well by what Larkin has described as the simultaneous technical and poetic dimensions of infrastructural systems; the gap, that is, between what infrastructure promises and what it can sustain. Deathbot platforms depend on heterogeneous and distributed technical arrangements such as cloud storage, API architectures, machine learning pipelines, and platform-specific data governance regimes. None of these is stable or insulated from the economic and organisational pressures that govern all digital infrastructure (Starosielski). This infrastructural fragility has direct consequences for the claim to immortality. The survival of a deathbot is a commercial arrangement, dependent on subscription models, platform viability, and the continued willingness of companies to maintain services for which the user base is, by definition, unable to advocate (Öhman and Floridi). The history of digital platforms is substantially a history of discontinuation and data loss; services shut down, interfaces change, APIs are deprecated, and the data that underpinned them become inaccessible or are deleted entirely (Kneese). Techno-solutionism consistently fails in the context of platforms, and there is no structural reason why deathbot platforms should be exempt from these pressures (Kneese). What is marketed as immortality is, in infrastructural terms, a service agreement, and the conditions of that agreement are neither transparent nor permanent. Beyond the question of platform viability lies a deeper infrastructural problem concerning the nature of what data systems capture and how they recombine it. The traces from which deathbots are constructed—messages, voice recordings, social media posts, browsing histories—are not neutral or comprehensive records of a person but partial, context-dependent, and platform-shaped outputs of specific systems with specific logics of capture and retention (van Dijck; Cheney-Lippold). What any given platform retains is determined not by what is most representative of a person but by what its architecture is designed to capture, which means that the dataset from which a deathbot is trained is already shaped by commercial, technical, and algorithmic decisions that have nothing to do with the person whose immortality is at stake. The recombination of these traces through machine learning inference (and with the data traces of others) compounds rather than resolves this problem: what is produced is not a reconstruction of the person but a probabilistic model trained on their digital residue, generating outputs that are statistically plausible given the training data but that bear an unstable relationship to what the person would actually have said, thought, or done (Bender et al.). There is, moreover, a political dimension to this infrastructural analysis that the language of immortality tends to obscure. Digital remains, in a status analogous in important respects to physical remains (Öhman), are not symmetrically or democratically distributed. Whose data are retained, at what volume, and in what condition is shaped by inequalities of access, platform use, and governance that reproduce, in the afterlife of data, the inequalities of life (Öhman; Recuber). The scale at which platforms now hold the digital remains of the dead does not translate into equivalent conditions of individual continuity; on the contrary, it produces the structural asymmetry this article has identified as central to understanding what digital immortality actually means in practice. Whoever controls these archives, as Öhman argues, effectively controls future access to a collective digital past, and this power has consequences also for the broader politics of memory and historical record. Data persist massively and indifferently, while individuals persist partially and only for as long as the commercial and technical conditions that sustain their particular configuration remain in place. It is this asymmetry that the concept of digital immortality, as deployed in deathbot discourse, systematically fails to acknowledge, and it is here that the consent and agency dimensions of the argument become most pressing. In life, as previously argued, the assembly of the digital self involves the ongoing participation of the person; the posting, deleting, contextualising, and curating through which data are produced and managed. This participation, however partial and constrained by platform architectures, retains at least the possibility of correction or withdrawal. Death removes that possibility entirely, leaving the traces to be assembled and reassembled by systems and commercial actors whose interests are not those of the person whose data they hold (Kasket). The deathbot does not merely inherit a static archive; it operationalises that archive in ways the person could not have anticipated and did not consent to (Stokes), under conditions of infrastructure and inference that were not available at the time the data were produced, and in a context of commercial reanimation that transforms the meaning and function of data that were often generated for entirely different purposes. The reframing of death as a technical problem amenable to data-driven solutions, which underwrites the entire deathbot enterprise, demands critical scrutiny precisely because it naturalises these commercial and infrastructural arrangements as care rather than extraction (Kneese). The Limits of Digital Immortality What deathbots reveal, when examined through the ontological and infrastructural conditions this article has traced, is not the achievement of immortality but its persistent and structurally necessary deferral. The immortality they claim is real as a cultural formation and as a commercial proposition draws on deep and enduring human desires and aligns itself with an ideological tradition. It produces, at the interface level, encounters that are genuinely felt and emotionally significant, but it is underwritten by conditions that ensure it cannot be delivered in any of the senses that the concept has historically carried. Deathbots promise a form of resurrection but deliver something closer to legacy (Cave), and even that legacy is partial, commercially contingent, and shaped by the logics of systems that were never designed with the integrity of the person in mind. The distance between what is claimed and what is delivered is not incidental to deathbot culture but constitutive of it, and it is not a distance that better technology or more comprehensive data capture will meaningfully close. The unified, recoverable self that the idea of immortality presupposes never existed in any robust sense, digitally or otherwise, and the question of whether deathbots can deliver it is, in this respect, secondary to the more fundamental recognition that what they are attempting to deliver was never available to be delivered in the first place. Rather than the person, or even a coherent representation of the person, what survives death in these systems is a dispersed set of digital remains that are partial, platform-shaped, and subject to the same forces of obsolescence and discontinuation that govern all digital cultural forms (Kneese; Öhman). The self that deathbots wish to reanimate was never, as this article has argued, the kind of unified and recoverable thing that reanimation in any meaningful sense requires; it is assembled and reassembled through ongoing processes of interaction that death forecloses. To reframe deathbots as a critical case study in digital persistence rather than a technology of immortality is not to dismiss the genuine and complex experiences of those who use them. We have attempted to unpack how the infrastructural and ontological conditions under which those encounters are made possible carry consequences that the language of immortality obscures, and to make the case that a more honest account of what digital systems can and cannot sustain is intellectually and ethically needed. The self does not persist, but neither can it fully disappear as the traces of the dead haunt the platforms that claim to have preserved them (Derrida, Specters of Marx; Fisher). References Bender, Emily M., et al. "On the Dangers of Stochastic Parrots: Can Language Models Be Too Big? 🦜" Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency, 2021. 610–23. . Brescó de Luna, Ignacio, and Belén Jiménez-Alonso. “Deathbots: Discussing the Use of Artificial Intelligence in Grief.” Studies in Psychology 45.1 (2024): 103–22. Cave, Stephen. Immortality: The Quest to Live Forever and How It Drives Civilisation. New York: Crown, 2012. Cheney-Lippold, John. We Are Data: Algorithms and the Making of Our Digital Selves. NYU P, 2017. Derrida, Jacques. Of Grammatology. Trans. Gayatri Chakravorty Spivak. Baltimore: Johns Hopkins UP, 1976. Derrida, Jacques. Specters of Marx: The State of the Debt, the Work of Mourning and the New International. Trans. Peggy Kamuf. New York: Routledge, 1994. Fabry, Renée E., and Mark Alfano. "The Affective Scaffolding of Grief in the Digital Age: The Case of Deathbots." Topoi 43.3 (2024): 757–69. Fisher, Mark. Ghosts of My Life: Writings on Depression, Hauntology and Lost Futures. Winchester: Zero Books, 2014. Floridi, Luciano, ed. The Onlife Manifesto. Springer, 2015. . Hollanek, Tomasz, and Karolina Nowaczyk-Basińska. "Griefbots, Deadbots, Postmortem Avatars: On Responsible Applications of Generative AI in the Digital Afterlife Industry." Philosophy & Technology 37 (2024): article 63. Kasket, Elaine. All the Ghosts in the Machine: Illusions of Immortality in the Digital Age. London: Robinson, 2019. Kidd, Jenny, and Eva Nieto Mcavoy. "Deep Nostalgia: Remediated Memory, Algorithmic Nostalgia and Technological Ambivalence." Convergence 29.3 (2023): 620-40. Kidd, Jenny, and Eva Nieto McAvoy. "Synthetic Afterlives: Deathbots as Affective Infrastructures of Memory." Memory, Mind & Media 4 (2025): e16. . Kidd, Jenny, and Eva Nieto McAvoy. AI Afterlives: Digital Memory and Synthetic Pasts. Bloomsbury Academic, 2026. Kneese, Tamara. Death Glitch: How Techno-Solutionism Fails Us in This Life and Beyond. New Haven: Yale UP, 2023. Kurzweil, Ray. The Singularity Is Near: When Humans Transcend Biology. New York: Viking, 2005. Larkin, Brian. “The Politics and Poetics of Infrastructure.” Annual Review of Anthropology 42 (2013): 327–43. Lifton, Robert Jay. The Broken Connection: On Death and the Continuity of Life. New York: Simon & Schuster, 1979. Morozov, Evgeny. To Save Everything, Click Here: The Folly of Technological Solutionism. New York: PublicAffairs, 2013. Morris, Meredith Ringel, and Jed R. Brubaker. "Generative Ghosts: Anticipating Benefits and Risks of AI Afterlives.” Proceedings of the 2025 CHI Conference on Human Factors in Computing Systems. New York: ACM, 2025. 1–14. Öhman, Carl. The Afterlife of Data: What Happens to Your Information When You Die and Why You Should Care. Chicago: U of Chicago P, 2024. Öhman, Carl, and Luciano Floridi. “The Political Economy of Death in the Age of Information: A Critical Approach to the Digital Afterlife Industry.” Minds and Machines 27.4 (2017): 639–62. Öhman, Carl, and David Watson. “Are the Dead Taking Over Facebook? A Big Data Approach to the Future of Death Online.” Big Data & Society 6.1 (2019). Papacharissi, Zizi, ed. A Networked Self: Identity, Community, and Culture on Social Network Sites. Routledge, 2010. . Recuber, Timothy. The Digital Departed: How We Face Death, Commemorate Life, and Chase Virtual Immortality. New York: NYU P, 2023. Sherlock, Alexandra. "Larger than Life: Digital Resurrection and the Re-Enchantment of Society." The Information Society 29.3 (2013): 164–76. . Starosielski, Nicole. "‘Warning: Do Not Dig’: Negotiating the Visibility of Critical Infrastructures." Journal of Visual Culture 11.1 (2012): 38–57. . Stokes, Patrick. Digital Souls: A Philosophy of Online Death. London: Bloomsbury Academic, 2021. Turkle, Sherry. Life on the Screen: Identity in the Age of the Internet. Simon & Schuster, 1995. Van Dijck, José. Mediated Memories in the Digital Age. Stanford: Stanford UP, 2007.
Zitieren
APA
McAvoy, E. N., Kidd, J., & Jones, B. (2026). The Limits of Digital Immortality. M/C Journal, 29(3). https://doi.org/10.5204/mcj.3285
Vancouver
McAvoy EN, Kidd J, Jones B. The Limits of Digital Immortality. M/C Journal. 2026;29(3). doi:10.5204/mcj.3285
BibTeX
@article{McAvoy2026,
author = {McAvoy, Eva Nieto and Kidd, Jenny and Jones, Bethan},
title = {The Limits of Digital Immortality},
journal = {M/C Journal},
year = {2026},
volume = {29},
number = {3},
doi = {10.5204/mcj.3285}
}
EUVIMED