Two Logics of Memory
M/C Journal · 18.06.2026 · 29 (3)
DOI: 10.5204/mcj.3280 ↗Kurz gefasst
Fachartikel in der Zeitschrift M/C Journal, online erschienen am 18.06.2026. Verfasst von María Josefina Alippi. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0).
Abstract (Originalsprache)
Introduction The concepts of “digital afterlife” and “digital legacy” are currently spreading across the Internet, signalling an emerging industry fuelled by artificial intelligence (AI) focussed on grief and loss. “Griefbots”, “deadbots”, and “digital twins”, derivatives of the increasingly widespread companion AI (Malfacini), offer replicas of deceased people by enmeshing generative artificial intelligence (Large Language Models of Transformer architecture) and surveillance-level data collection. The self-defined “grief tech” offers to keep alive the memory of those who have passed through the creation of digital representations with conversational interfaces with which the user can indefinitely talk, mimicking appearance and behaviour through the use of gathered online data—such as voice and video recordings, biometrics, photographs, emails and social media interactions. These products ultimately offer a restoration of presence before loss and the mitigation of grief through digital performance. We argue that underpinning these AI grief technologies is an understanding of what memory is that operates through a conflation of computational processes and human memory. Furthermore, we contend that this equation of memory with a set of computational procedures of high complexity underlies the commodification of human interior life within an expanding political economy that seeks to mine and capitalise on emotional resources through technology; an operation embedded in the rapid integration of AI as a new medium, which ultimately works against the human faculty of memory as an essential (re)constitutive narrative process of the self. The Underlying Conflation of Memory and Computation Several startups have emerged during the past years offering digital products based on providing simulations of deceased people. The names of these companies already advance absolute claims: Replika is directly named after the concept of a replica of a person; You, Only Virtual speaks directly to a possibility of retaining selfhood in a split version; HereAfterAI obliquely names death itself. Their marketing pitches offer to help the bereaved to remember their loved ones, granting a remedy to forgetfulness—HereAfterAI, for example, promises to “reinvent remembering”, while You, Only Virtual claims to “transform how the world experiences grief and remembrance”. Bolder versions of these products offer straightforward communication with the deceased: You, Only Virtual offers on its front page “AI-powered communication with those we’ve lost”; similarly, Seance grants to “let your loved ones speak to your heart once again”. A long-standing science fiction trope underlies these marketing promises: the imaginary of uploading, downloading, and retrieving consciousness from its physical embodiment in a person to an external device that can contain what ultimately constitutes the self. The dystopian 2019 series Years and Years depicts a near future in which a person’s mind is uploaded onto a server, to be retrieved by subsequent generations as though consulting an interactive digital archive in a largely disembodied world in which we all interact as holograms—therefore allowing for the complete indifferentiation between actual (disembodied) people and those who no longer live but have had their minds uploaded. The 2021 series Solos moves the premise further and closer to the present: a man, terminally ill, purchases an identical fleshed-out replica of himself, designed to replace him as a father to his children after his death. The man tests and confirms that his replica thinks and feels—and would react in future hypothetical scenarios he has not yet lived—exactly as he would. As Carl Freedman argues, what is distinctive to the science fiction genre is that it keeps a rationally accountable link, by means of which its premise never falls onto fantasy ground but stays in the realm of reality, managing to stay within its parameters through explanations provided by—fictional—science. In the Solos episode, the replica is able to come up with its own spontaneous reasoning—not executing mere archival retrieval, but generative responses. Generative AI systems’ ability to produce seemingly original content from what the Internet’s archive has given them, largely due to their extraordinarily plastic use of language, has already grounded that imaginary link. The recently popularised notion of “neural networks” takes that further, as it circulates beyond the AI study field, advancing imaginaries of the replication of the brain’s functioning, whereby the ability to recreate neurons would imply the ability to recreate the way a brain works. In all these accounts of extended presence through synthetic forms of conservation of the self, an interplay is put into motion between replication and actual identity of the replicated being. Whether retained as a form of soul or consciousness in an external device (as depicted in Years and Years) or replicated as a simulation that manages to bring up the actual person by means of sheer algorithmic performance (as in the Solos episode), a core conceptualisation of selfhood is at stake. Insofar as these depictions foster the premise that someone can be replicated through a management of their gathered data, it incites an idea of self and personhood that can in some way be eventually equated to a sum of data; a notion which can ultimately be traced back to the computational model of memory. Two Logics of Memory The longstanding brain as computer metaphor, whose scientific roots lie in mid-twentieth-century cybernetics and cognitive science as advanced by cybernetics founders Norbert Wiener and John von Neumann (Hayles), migrated from metaphorical expressions stemming from the computational sciences into everyday language as an uncontested analogy whose figurative nature is often overlooked (Innocenti). As scholars have persistently pointed out throughout the past decades (Carbonell et al.; Colburn and Shute; Möck), in technology, particularly, the metaphors designed to circulate complex and abstract concepts usually have a bidirectional impact, ultimately shaping the very research questions that inform objects, concepts and entire fields of study (Gozzi). On this basis, the above-mentioned imaginaries of fiction as well as the grief-oriented companies’ marketing rhetoric resort to a conflating conceptualisation of memory stemming from the brain as a computer/computer as a brain metaphor: human memory is equated to a computer’s technical capacity for storage and retrieval of information, whether as working memory actively processed during operation or as long-term storage written to disk. Yet the differences are vast and substantial. Computational memory is the technical capacity of digital systems to store, retrieve, and transfer discrete units of data. It is the mechanism by which a machine records and accesses information, operating through fixed electronic circuits and binary code. It is static in nature, given that stored information remains exactly identical and at the same location. Neither time nor context alters it. Human memory operates according to an entirely different logic. In his account of the biology of memory, Nicholas Carr provides a historically grounded basis that debunks the metaphor of the brain working as a storage system. While computer memory is static and binary, biological memory is alive and unavoidably embodied. Citing neuroscientist Kobi Rosenblum, Carr explains that while a computer absorbs information and immediately saves it in fixed form, the human brain continues to process information long after it is received, and the quality of memories depends on how that information is processed over time. Unlike computational memory, which is immediate and every time identical to its input, memory consolidation is a slow, organic and multi-stage biological process. Short-term memories become long-term ones through complex biological processes of protein synthesis, growth of new synaptic terminals, and biochemical and anatomical changes at the cellular level. The quality of memories depends not on the quality of input but on how that information is processed across time, through sleep, dreams, pleasure, and trauma. Far from being a recollection of past events, memory is an active, organic mesh of different recollections and dispositions that consolidate into a set of representations always open to being re-signified. This biological understanding finds a psychic correlate in Freud. In Civilization and Its Discontents, opening with the famous analogy of Rome as a psychic space—a city in which every prior stage of construction coexists simultaneously with every subsequent layer, though transformed throughout time—Freud affirms that in the mind, unlike in physical space, “nothing that has once taken shape can be lost”, and that “the retention of all previous stages, together with the final shape, is possible only in the mind” (22). This does not mean that the human mind is not able to forget; rather, it argues that every lived experience persists in the psychic apparatus operating at varying levels of depth, reciprocally shaping and being re-shaped by the experiences that come after it. What persists in the mind does so dynamically, and appears transformed, repressed, distorted, only partially accessible, and always mediated by the ongoing psychic life of the subject. From a phenomenological perspective, this active memory is further elaborated by Ricoeur, for whom memory is an effortful practice of self-constitution. The self is the ongoing narrative it constructs across time—as opposed to the mechanical addition of the archive, the sum of what it has recorded. This narrative identity (what Ricoeur calls ipse) is constituted through the living act of remembering and interpreting, not stored within it; it is what ceases with death and cannot be archived. Memory, Ricoeur further argues, is crucially constituted through forgetting. We do not remember everything, and it is through that irreducible selectivity that meaning is made. In this sense, a disembodied memory that can be placed somewhere other than the body incites a controversy regarding the possibility of communicating with a loved one “again”. Whatever the fidelity of the replica, deadbot, or digital twin, what they promise to bring into presence to counter the absence and the pain of loss is not the person that is no longer here, but something else. The Work of Mourning, Disrupted In Mourning and Melancholia, Freud defines mourning as work (Trauerarbeit, the work of grief), a necessary interior labour through which the ego gradually withdraws its libidinal attachment to the lost object (in the case of loss, the deceased loved one) and becomes able to reorient this libido towards the world. Mourning is not forgetting; it is an active—and painful—metabolisation of loss, through which the self is reconstituted in the absence of what it loved. Freud calls melancholia what occurs when this work fails, as the ego refuses to relinquish the lost object and turns against itself. What distinguishes the two is the actual doing of the mourning work. In the scenario offered by AI-driven griefbots, digital twins, and deadbots, the bereaved user is not invited to relinquish the lost object but to retain it, sustaining the illusion of its presence. The extent to which this might affect the grieving process is something to which psychoanalysis is starting to attend. As Lemma points out, “griefbots raise important psychological and ethical questions about whether machine-tethered mourning is helpful to the individual and to society or whether it is even accurate to refer to it as ‘mourning’ since its function may well turn out to be an evasion of mourning”. The question is raised as to whether “grief technology”, rather than facilitating mourning, might tamper with its very realisation. Insofar as grief constitutes, from a psychoanalytical perspective, “a lifelong process of taking the other into the self and bearing the pain of loss” (Lemma, 551), the possibility exists that, far from providing what can be understood as a tool for dealing with loss, they embody a novel narcotic practice—as Freud hypothesised in Civilization and Its Discontents regarding the various ways in which humans attempt to counter the pain intrinsic to human life—which might induce only a temporary freedom from the hardships of grief, while disrupting the transformative work of the mourning process for some of its users, were it to circulate as a freely available product for purchase. Despite considering potential healing uses for griefbots, the attempt to digitally immortalise the dead, as Lemma states, “can only be viewed suspiciously because it potentially works against the grain of loosening the ties that bind us to the dead” (550). In Understanding Media, Marshall McLuhan argues that the adoption of a new medium entails an extension of human capacities that progressively takes over, eventually amputating the ability for the skills now performed by the new technology—just as the wheel extends the foot and diminishes the walker. To extend one faculty through a technical prosthetic is to numb and atrophy the capacity it replaces: “the nature of the medium, of any and all media, [is] in the true Narcissus style of one hypnotized by the amputation and extension of his own being in a new technical form” (McLuhan 22). The prosthetic, in relieving the body of effort, relieves it also of use, and what ceases to be used ceases, in time, to function. The AI products oriented to grief present themselves as extensions of the relational bond with the deceased, promising to keep alive what would otherwise fade, through the premise of a form of memory that equates the human to complex computational and algorithmic processes. The prospective effects of this technology are not to be dismissed straightforwardly; media have always reshaped collective and individual practices, and new conceptualisations of human phenomena, including death, have always followed. What is at stake here is the particular political economy driving this reshaping. Platform Capitalism and the Political Economy of “Grief Tech” The emergence of these technological products targeting grief cannot be separated from the political economy that sustains them. As Evgeny Morozov has argued, the solutionist logic pervading Silicon Valley systematically reframes complex human problems as technical deficiencies awaiting a product fix. Grief is easily understood, within this framework, as a malfunction—a broken state against an ideal of a fully functioning, pain-free being—rather than the necessary and irreplaceable process through which a human being incorporates loss and continues to become itself. The growing body of scholarship on the commodification of digital death (Kneese; Savin-Baden and Mason-Robbie) has begun to map the contours of this reframing, tracing how dying and mourning have become addressable markets. Srnicek's analysis of platform capitalism identifies data extraction as the foundational logic of the digital economy: human behaviour, preference, interaction, and affect are the raw material of value. Technologies aimed at grief represent the most intimate frontier of this extraction. It is not coincidental that two of the largest players in the sector, Meta and Microsoft, have already secured patents for post-mortem user simulations: the deceased user, archived, is a dataset that continues to generate behavioural data and, most importantly, continued emotional engagement from the living. The bereaved do not merely consume a product: they become, in the act of grieving through a platform, a source of further data, further emotional profiling, and further capture. The commodification of inner life here might reach an extreme form as it focusses on one of the most vulnerable states of human beings to monetise subscription services. At its core, grief-oriented technologies present one of the sharpest instances of what is at stake in the unrestricted integration of AI into everyday life: the gradual substitution of computational logics for human ones in domains where that substitution carries irreversible psychic consequences. The two logics of memory this article has traced underpin two different answers to what it means to be human: to lose, and to go on living. Interrogating the assumptions these products advance is, ultimately, a matter of what kind of subjectivity a society chooses to protect. References Carbonell, Javier, et al. “The Role of Metaphors in the Development of Technologies: The Case of the Artificial Intelligence.” Futures 84 (2016): 145–53. Carr, Nicholas. The Shallows: What the Internet Is Doing to Our Brains. New York: W. W. Norton, 2010. Colburn, Timothy R., and Gary M. Shute. “Metaphor in Computer Science.” Journal of Applied Logic 6.4 (2008): 526–33. Freedman, Carl. Critical Theory and Science Fiction. Middletown: Wesleyan UP, 2000 Freud, Sigmund. Civilization and Its Discontents. Trans. James Strachey. New York: Norton, 1961 [1930]. ———. “Mourning and Melancholia.” The Standard Edition of the Complete Psychological Works of Sigmund Freud. Vol. XIV. London: Hogarth Press, 1957 [1917]. 237–58. Gozzi, Raymond, Jr. The Power of Metaphor in the Age of Electronic Media. Cresskill: Hampton Press, 1999. Hayles, N. Katherine. How We Became Posthuman: Virtual Bodies in Cybernetics, Literature, and Informatics. Chicago: U of Chicago P, 1999. Innocenti, Giorgio M. “Neural Computing: The Metaphorical, Cultural Roots of Brain Models.” Culture and Brain 1.1 (2013): 64–75. Kneese, Tamara. Death Glitch: How Techno-Solutionism Fails Us in This Life and Beyond. New Haven: Yale UP, 2023. Lemma, Alessandra. “Mourning, Melancholia and Machines: An Applied Psychoanalytic Investigation of Mourning in the Age of Griefbots.” International Journal of Psychoanalysis 105.4 (2024): 542–63. Malfacini, Kim. “The Impacts of Companion AI on Human Relationships: Risks, Benefits, and Design Considerations.” AI & Society 40.7 (2025): 5527–40. McLuhan, Marshall. Understanding Media: The Extensions of Man. New York: McGraw-Hill, 1964. Möck, Leonie A. “Prediction Promises: Towards a Metaphorology of Artificial Intelligence.” Journal of Aesthetics and Phenomenology 9.2 (2022): 119–39. Morozov, Evgeny. To Save Everything, Click Here: The Folly of Technological Solutionism. New York: PublicAffairs, 2013. Ricoeur, Paul. Memory, History, Forgetting. Trans. Kathleen Blamey and David Pellauer. Chicago: U of Chicago P, 2004. ———. Oneself as Another. Trans. Kathleen Blamey. Chicago: U of Chicago P, 1992. Savin-Baden, Maggi, and Victoria Mason-Robbie, eds. Digital Afterlife: Death Matters in a Digital Age. London: Routledge, 2020. Srnicek, Nick. Platform Capitalism. Cambridge: Polity, 2016.
Zitieren
APA
Alippi, M. J. (2026). Two Logics of Memory. M/C Journal, 29(3). https://doi.org/10.5204/mcj.3280
Vancouver
Alippi MJ. Two Logics of Memory. M/C Journal. 2026;29(3). doi:10.5204/mcj.3280
BibTeX
@article{Alippi2026,
author = {Alippi, María Josefina},
title = {Two Logics of Memory},
journal = {M/C Journal},
year = {2026},
volume = {29},
number = {3},
doi = {10.5204/mcj.3280}
}
EUVIMED