Technonecromancy: Simulacra of Presence and the Politics of Death in the Age of Generative Artificial Intelligence*

Tecnonecromancia: simulacros de presencia y políticas de la muerte en la era de la inteligencia artificial generativa

DOI10.22430/21457778.3723 Logotecnologicas PDF Table Figure

Received: October 17, 2025
Accepted: May 5, 2026

How to reference / Cómo referenciar
da Silva, G. S., & Ulbricht, V. R. (2026). Technonecromancy: Simulacra of Presence and the Politics of Death in the Age of Generative Artificial Intelligence. Trilogía Ciencia Tecnología Sociedad , 18(39), e3723. https://doi.org/10.22430/21457778.3723

 

Abstract

This article theorizes technonecromancy as a novel regime of mediation that operationalizes spectral presence through generative models, voice cloning, and synthetic video, shifting the relation to the dead from archival recollection to dialogic simulation. Conceptually, it integrates science, technology, and society studies with phenomenology and 4E cognition to argue that posthumous agents are phenomenologically sufficient to reorganize mourning yet ontologically insufficient as continuants of persons. Empirically, it reads cultural and commercial cases (such as griefbots and animated photo apps) as epistemologies of the technocultural present, demonstrating how engineered responsiveness can both scaffold continuing bonds and entrench dependence, thereby reconfiguring rituals of loss and memory. Normatively, it advances a consent-first, provenance-rich governance program, with mandatory disclosure, machine-readable labeling, and revocability, aligned with frameworks, such as the European Union Artificial Intelligence Act’s transparency duties for synthetic content, and calls for fiduciary stewardship of digital remains. Thus, this study situates these proposals within the emerging digital afterlife industry, noting that social acceptability collapses absent explicit pre-mortem authorization and that professionalization is beginning to reshape industry standards. The result is a political-economic diagnosis: technonecromancy is a world-making apparatus whose risks and benefits can only be adjudicated where ontological critique, clinical harm minimization, and market governance converge.

Keywords: digital afterlife, governance, necrocapitalism, STS studies, technonecromancy.

Resumen

Este artículo teoriza la tecnonecromancia como un novedoso régimen de mediación que operacionaliza la presencia espectral a través de modelos generativos, clonación de voz y video sintético, desplazando la relación con los muertos desde el recuerdo archivístico hacia la simulación dialógica. Conceptualmente, integra los estudios de ciencia, tecnología y sociedad con la fenomenología y la cognición 4E para argumentar que los agentes póstumos son fenomenológicamente suficientes para reorganizar el duelo, pero ontológicamente insuficientes como continuantes de las personas. Empíricamente, analiza casos culturales y comerciales (como los griefbots y las aplicaciones de fotos animadas) como epistemologías del presente tecnocultural, demostrando cómo la capacidad de respuesta diseñada puede tanto andamiar los vínculos continuos como afianzar la dependencia, reconfigurando así los rituales de pérdida y memoria. Normativamente, propone un programa de gobernanza basado en el consentimiento previo y rico en procedencia, con divulgación obligatoria, etiquetado legible por máquina y revocabilidad, alineado con marcos como los deberes de transparencia de la Ley de Inteligencia Artificial de la Unión Europea para el contenido sintético, y aboga por una custodia fiduciaria de los restos digitales. Así pues, esta investigación sitúa estas propuestas dentro de la emergente industria de la vida digital post mortem, señalando que la aceptabilidad social colapsa en ausencia de una autorización explícita pre mortem y que la profesionalización está comenzando a remodelar los estándares de la industria. El resultado es un diagnóstico político-económico: la tecnonecromancia es un aparato de construcción de mundos cuyos riesgos y beneficios solo pueden ser adjudicados donde convergen la crítica ontológica, la minimización de daños clínicos y la gobernanza del mercado.

Palabras clave: vida digital post mortem, gobernanza, necrocapitalismo, estudios cts, tecnonecromancia.

INTRODUCTION

The emergence of posthumous simulations (interactive digital entities generated through large-scale data extraction, deep learning architectures, and algorithmic modeling of personality and voice) has inaugurated what may be termed a thanatotechnical shift in the history of mediation. Against this backdrop, the contemporary condition of technoscience manifests itself as an ontological experiment that dissolves the boundaries between the living and the non-living, the animate and the inanimate, the mortal and the machinic. Technology ceases simply to preserve traces of the dead and begins instead to perform their semblance, animating new forms of presence-in-absence that transform the ontology of memory and mourning (Kidd & Nieto McAvoy, 2023; Kopelman & Frosh, 2025; Rodríguez Reséndiz & Rodríguez Reséndiz, 2024).

Traditional digital memorialization, such as social media archives, photo repositories, or textual legacies, presupposed a static metaphysics of trace. However, the rise of Generative Artificial Intelligence (GAI) introduces a dynamic metaphysics of simulation—a capacity to produce responsive avatars, voice clones, and conversational agents that replicate both the appearance and the behavioral potentialities of the deceased (Cheng, 2025; Lei et al., 2025). These emergent entities complicate the philosophical distinction between representation and reanimation. In turn, this situation raises the question of whether the algorithmic reproduction of the dead signifies an extension of the archive or a new ontological regime altogether, a regime where virtuality replaces and exceeds reality (Baudrillard, 1994; Hansen, 2006).

Within the field of Science, Technology, and Society (STS) studies, death and digitality have rarely been theorized beyond the framework of grief technologies (grief techs) and digital afterlife services (Khusyainov & Urusova, 2024; Reese, 2023). While the existing literature is highly valuable, it remains largely descriptive and empiricist, addressing ethical regulation, data protection, or therapeutic affordances. What remains absent is a rigorous ontological and epistemological investigation into how technological mimesis itself reorganizes the existential coordinates of finitude, presence, and selfhood. In other words, STS studies on death have yet to fully confront its own metaphysical implications. This article seeks to start to fill that lacuna.

Building on adjacent denominations, such as “pseudo-necromancy” (Corcoran & Di Carlo, 2022), “digital necromancy” (Gruvæus, 2023; Morse, 2024), and “cyber-necromancy”, this study adopts technonecromancy to designate the technically mediated conjuration of the dead for interactive simulation. This term is suggested for three reasons. First, pseudo-necromancy foregrounds the rhetoric of falsity implied by pseudo, which prejudges the phenomenon as mere sham rather than an operative regime of mediation (Corcoran & Di Carlo, 2022). Second, digital necromancy narrows the field to a digital substrate, whereas the emphasis of this study is on technics in the broad sense, encompassing devices, procedures, infrastructures, and institutional protocols. These are capable of analog, hybrid, and computational instantiations, even if contemporary cases are overwhelmingly digital (Gruvæus, 2023; Morse, 2024). Third, cyber-necromancy over-indexes the locus of cyberspace, occluding off-platform, embodied, and juridical-economic mediations such as studio pipelines, voice-cloning hardware, and estate contracts that co-produce the effect. The techno prefix thus marks the convergent assemblage that automates communion with the dead, while necromancy retains the cultural semantics of summoning. As an analytical and diagnostic category, technonecromancy captures a shift from archival presence (the storage of traces) to performative presence (the algorithmic, often, but not necessarily, digital simulation of relationality). Consequently, this concept clarifies how contemporary infrastructures reconfigure the phenomenology of death, the ethics of consent, and the semiotics of memory.

From a philosophical standpoint, technonecromancy invites a re-examination of the ontological status of simulation: whether these posthumous agents are extensions of an individual’s digital remains, derivative semiotic entities, or entirely new synthetic beings. It thus traverses a conceptual terrain spanning the phenomenology of embodiment (Gallagher, 2014; Moran, 2013; Varela et al., 2017), the metaphysics of identity (Nicolelis, 2020; Oderberg, 1993; Parfit, 1987), and the sociology of algorithmic mediation (Couldry & Hepp, 2018). Through this interdisciplinary lens, this field is construed as a techno-ontological dispositif—one of memory, mourning, and control, wherein the digital economy subsumes even mortality itself into its circuits of reproduction and exchange (Attali, 2003; Zuboff, 2019).

Accordingly, the following research question orients this inquiry: How do mimetic technologies reconfigure the boundary between life, death, and presence? To answer this, the paper integrates conceptual analysis, cultural case studies, and philosophical genealogy. Its objective is threefold: (i) to define and formalize technonecromancy as a conceptual construct within STS studies; (ii) to map its enabling technologies, including deepfakes, voice cloning, Large Language Models (LLMS), as components of a convergent generative assemblage; and (iii) to interrogate the ethical, ontological, and political consequences of this assemblage.

The contribution of this study is both conceptual and normative. Conceptually, it provides a vocabulary for articulating the emergent symbiosis between GAI and mortuary culture, situating this phenomenon within the broader logic of technocapitalist reproduction. Normatively, it calls for a critical ethics of posthumous mediation that reclaims dignity, finitude, and mourning from the neoliberal commodification of presence. The argument proceeds from the premise that, in a culture where even death becomes generative, the limits of the human are extended and algorithmically overwritten. This process, sketched in Figure 1, describes how the self is converted into a perpetually updatable simulation.

Figure 1. Demand curves of the Colombian electric sector used in the simulation.
Source: Authors.

In this sense, technonecromancy both describes a technological innovation and names a paradigmatic mutation in the ontology of subjectivity. It is both a symptom and an operator of a civilization that seeks to abolish loss through computation, transmuting the metaphysical drama of mortality into an aesthetic of recursive persistence. The ensuing analysis, therefore, engages the ethics of data necropolitics (Mbembe, 2003), the hauntological temporality of algorithmic memory (Derrida, 2024; Fisher, 2014), and the phenomenotechnics of simulation (Hansen, 2024; Hui, 2016), arguing that the technological resurrection of the dead marks the automation of mourning itself.

THEORETICAL FRAMEWORK

In order to articulate technonecromancy as a robust analytical operator, this section is anchored in two primary theoretical axes within STS studies: (i) the sociotechnics of death (digital afterlife systems and grief technocultures) and (ii) the ontology of simulation and hyperreality (the regime of the simulacrum). By establishing these two pillars, the infrastructural and ontological shifts brought by GAI are consolidated. To prevent theoretical overextension and maintain analytical hierarchy, other frameworks mobilized in this analysis—specifically phenomenology, 4E cognition, and necropolitical economy—are explicitly treated as auxiliary lenses. These secondary lenses are deployed strategically to decode specific psychological effects and market incentives, subordinating them to the primary STS studies critique of mediated posthumous presence.

Emergent grief techs, such as chatbots, voice clones, and posthumous avatars driven by Artificial Intelligence (AI), aim to soften the rupture of loss by keeping users connected to the deceased (Corcoran & Di Carlo, 2022; Morse, 2024). However, critics warn that these digital immortals, while keeping the dead "among the living," risk interfering with natural grieving processes (Reese, 2023). From an STS studies perspective, these are not mere artifacts but sociotechnical devices that configure desires, economic incentives, and affective labor, instantiating a thanato-infrastructure where death is remediated and monetized. This signals a critical shift from preserving archival traces to simulating interactive presence.

This simulated presence carries pathologies. Grief tech can become a "crutch," fostering dependence that entrenches denial rather than enabling closure (Gruvæus, 2023). Such technologies demand a critical genealogy of "mourning as a service," examining how they converge with issues of consent, data legacy, and affective capitalism (Voinea, 2024). Technonecromancy is thus a sociotechnics of death that generates relational agents, not just memorials, forcing engagement with deeper ontological and cognitive questions.

The ontological status of these simulated dead requires consulting simulation theory, particularly the work of Baudrillard, Derrida, Hansen, and Hui. Baudrillard’s (1994) doctrine of simulacra posits a transition into hyperreality, where signs no longer refer to reality but only to other signs, and the simulation ultimately eclipses the original (Danesi, 2024). Technonecromantic agents are not mere representations but hyperreal presences, operating through generative simulation to produce novel "afterlives" rather than simple mimicry (Hollanek & Nowaczyk-Basińska, 2024).

Derrida's (2024) hauntology adds a layer of complexity, framing the specter as an unsettling trace (neither alive nor dead) that resists closure. Posthumous agents function as technological specters, where Baudrillard’s collapse of referentiality meets Derrida’s interruption of temporality. Further, Hansen (2006) offers a mediating path, conceiving of digital simulations as "bodied code," where presence is always translated and deferred through procedural media. Hui (2016) extends Hansen’s stance, arguing that code is world-constitutive, instantiating ontological economies where death becomes a mutable relation, not a fixed boundary. In this framework, simulacra become agents and the spectral becomes constitutive of presence.

The algorithmic analogues of personhood challenge stable notions of identity. Parfit’s (1987) philosophy, which defines identity through psychological continuity (Relation R) rather than substance, allows for a divisible, branching self, questioning whether a simulation preserving discursive patterns is the "same person" (Afroza, 2023). However, this view is countered by embodied cognition theories (Varela et al., 2017), which assert that cognition arises from sensorimotor coupling with an environment. A simulation, while discursively rich, lacks this lived corporeality and experiential intentionality.

This ontological gulf is bridged by the human propensity to anthropomorphize algorithms, projecting agency and continuity onto the simulation, making it feel "almost alive" (Richter, 2025). This aligns with Human-Computer Interaction (HCI) research showing how interface affordances foster agency attribution (Nowak, 2004; Simas & Ulbricht, 2024), creating a paradox of authenticity where perceived realness is decoupled from ontological status. Technonecromancy thus occupies the tension between algorithmic mimicry and phenomenological difference, and between discursive continuity and experiential lack.

RESEARCH APPROACH

This inquiry is not empirical in the positivist or data-driven sense but resides instead in a theoretical-conceptual, critical, and exploratory register. Its posture is diagnostic, as opposed to confirmatory: the goal is to problematize and reframe latent conditions, rather than to test predetermined hypotheses. To that end, the method weaves conceptual analysis, genealogical historiography, and a phenomenotechnical reading of cultural artifacts and platforms into an integrated hermeneutic (Chimisso, 2008; Rheinberger, 2005).

From the outset, the orientation is exploratory-conceptual: technonecromancy is a nascent domain, requiring thick articulation rather than empirical generalization. The researcher assumes the role of conceptual cartographer, intervening in the discursive field, positing provisional formulations, testing them in cultural reference points, and refining them in iterative cycles. In doing so, the approach remains abductive, moving from concept to case, through refinement, and back to reapplication, with the aim not of closure but of progressively disciplined specification. Genealogical historiography, informed by Foucault’s “history of the present” (Garland, 2014; Rockhill, 2020), is deployed not to locate an origin but to trace the contingent technological, institutional, affective, and discursive assemblages that made the notion of algorithmic necropresence thinkable. By mapping ruptures, sedimentations, and silenced alternatives, genealogy reveals how infrastructures of memory, data capitalism, mortuary imaginaries, and law co-emerged to render technonecromantic operations plausible.

Complementing genealogy, the phenomenotechnical dimension reads technological artifacts, including deepfakes, avatar systems, and chatbots, as mediators of experience—media that structure temporalities, affordances, lags, failures, and spectral tensions (Hui, 2016; Keightley, 2013). This reading attends to how glitches, delays, interface cues, and algorithmic misalignment reveal the distance between simulation and presence; these traces become the locus where haunting is engineered.

To render this triadic methodology concrete, this article analyzes a purposive sample of cultural-technological case studies, namely the digital resurrection in Rogue One, Deep Nostalgia (MyHeritage), and HereAfter AI, alongside emerging griefbot platforms. These cases were selected based on three explicit criteria: (i) they represent distinct technical modalities of technonecromancy (cinematic representation, visual or facial animation, and conversational or interactive AI), (ii) they possess high public visibility and market penetration, and (iii) they have served as focal points for significant ethical and affective contestation in contemporary public discourse. Functioning as strategic semiotic probes, these cases serve as epistemological windows into how technonecromantic logics are negotiated in the public imaginary and the Digital Afterlife Industry (DAI).

To operationalize this analysis, the research examined a specific corpus of primary and secondary materials for each selected case. The analyzed sources include the platforms' terms of service and user agreements, promotional materials and marketing discourses, journalistic coverage, and publicly available user testimonials. The analytical procedure consisted of a critical interface hermeneutics, evaluating User Experience (UX) and User Interface (UI) affordances, system latency, and design choices, coupled with a thematic discourse analysis to identify how these platforms frame death, memory, and artificial presence. By triangulating these sources, the analysis surfaces the structural tensions, power effects, and ontological claims embedded in these technologies.

Reflexivity is structurally incorporated, with ongoing attention to the role of the analysis in shaping interpretations through triangulation across technical documentation, user reports, media discourse, and interface affordances. The limitations of this approach are acknowledged, including the lack of large-scale empirical validation, the opacity of proprietary systems, and the risk of analogical overreach. Nevertheless, this methodology functions as a discursive apparatus for the emergence of technonecromancy as a critical object in STS studies, laying the groundwork for subsequent empirical extension, normative elaboration, and technological critique.

ANALYSIS AND DISCUSSION

In this section, the triadic conceptual apparatus developed in the theoretical framework is applied to the selected cultural and technological cases in order to surface how technonecromancy reconfigures the architecture of mourning, identity, ontology, ethics, and economy. The analysis is organized across four interlocking subaxes: (i) mourning and AI-mediated attachment; (ii) simulated identity and embodied cognition; (iii) ontology of simulacra and hyperreality; and (iv) ethics, posthumous consent, memory commodification, and necrocapitalism. Each subsection engages in a dialectical movement, progressing from the technological morphology of the case through its existential implications to a normative interrogation.

Mourning and Artificial-Intelligence-Mediated Attachment

The promissory rhetoric of technonecromancy—that grieving subjects might sustain continuing bonds with the dead—draws knowingly on a decisive revision within bereavement studies. Against the twentieth-century orthodoxy that healthy mourning requires severance, Klass et al. (1996) repositioned the persistence of an inner relationship with the deceased as potentially adaptive rather than pathological. In their frame, the bereaved reorganizes a world saturated by the dead, not by forgetting, but by recontextualizing presence. Contemporary grief techs radicalize this revision by externalizing the bond: what was once an interiorized conversation becomes an algorithmically mediated dialogue, objectified in interfaces, avatars, and voice clones (Gruvæus, 2023; Morse, 2024). The very grammar of attachment is thus displaced from intrapsychic symbolization to interactive simulation, a shift that is sociotechnical as much as psychological.

At the center of this displacement is the griefbot, a computational agent trained on the digital footprints of the deceased that affords two-way conversational presence. Jiménez-Alonso and Brescó de Luna (2023) define it precisely as a conversational artifact that operates on legacy digital traces and enables communication with a synthesized version of the deceased, evoking a novel form of mediated connection. What is inaugurated here is a new means of remembering and a new mode of presence that is performative, responsive, and programmable, in which absence is permanently renegotiated through cycles of questions and answers, updates, and fine-tuning.

This externalization, however, imposes a phenomenological and clinical cost. Lindemann (2022) proposes that griefbots can block the transition from the "world-with-the-dead" to the "world-without-the-dead," which establishes an interactive loop that functionalizes denial and delays the process of mourning. From a normative perspective, the author advocates for explicit guardrails, framing such agents as technologies subject to protocols, with transparency regarding risk, governance, and usage limits. This argument converges with the literature on extended affect and "affective scaffolding," according to which technical-environmental arrangements can sustain—and also exacerbate—dependent emotional dynamics (Colombetti & Krueger, 2015). By offering infinite replies and contingent responsiveness, technonecromances introduce a temporal distortion (a present continuous) that unanchors mourning from its gradients of finitude and reinscribes it in routines of maintenance.

The affective micrology of interfaces like Deep Nostalgia evidences the engineering of this haunting. Kidd and Nieto McAvoy (2023) describe the conversion of static photos into animated clips through deep learning, in which blinks and minimal smiles catalyze oscillations between recognition and estrangement, an algorithmic nostalgia that updates memory through the simulation of gesture. Similarly, Kopelman and Frosh (2025) conceptualize the “algorithmic as if”: a regime of responsive appearance that suspends the ontology of the referent and performativizes the bond as if there were a returning presence. The result is a technological haunting in which movement is sufficient to summon affect but insufficient to resolve absence, an in-between that retains the subject in the loop of iteration.

From a clinical-nosographic perspective, this in-betweenness is not neutral. The formalization of prolonged grief disorder in the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5-TR) and its parallel discussion in the eleventh revision of the International Classification of Diseases provide a backdrop for reflecting on specific vulnerabilities (Eisma, 2023; First et al., 2023). By segmenting the time of mourning with interaction sessions and recurring check-ins, technonecromantic agents can override acceptance mechanisms, favoring rumination loops and emotional dependence, especially in at-risk subgroups. This does not mean demonizing mediation but recognizing ambivalence and categorizing outcomes according to psychosocial profiles and concrete designs, and this ambivalence already appears in public testimony. Recent reports document consoling applications such as voice cloning to "hear" a deceased parent, and family concerns, straining privacy, consent, and grief (Richter, 2025). The same article shows that the grief-tech ecosystem (StoryFile, HereAfter, Eternos) oscillates between codes of consent and promises of conversational legacy, cementing a public imagination in which the return of the dead is inscribed as a platform feature.

The question, then, shifts to the ethical design of mediation. Amershi et al. (2019) offer a consolidated set of principles, including making clear what the system can do, facilitating recovery, and managing changing capabilities. These guidelines can be adapted to grief tech through features such as gradual farewell modes, session limits and cooldowns, explicit simulation statements, deliberate decay of responsiveness over time, and revocation and archiving controls. Such heuristics should be co-validated by clinical psychology and ethics boards, especially when the product advertises itself as therapeutic-adjacent. To guide this coupling between clinical practice, design, and regulation, the synoptic framework in Table 1 is proposed, which translates evidence and guidelines into operational standards for grief tech.

Table 1. Risk-mitigation heuristics for mourning mediated by artificial intelligence
Risk vector (clinical or ethical)
Mechanism in AI mediation
Proposed mitigation (operational)
Temporal distortion and compulsive checkingInfinite responsiveness, novelty generationSession caps, cooldowns, usage nudges, graduated “goodbye mode”
Illusion of presence or ontological confusionAnthropomorphic cues, voice realism, latency smoothingPersistent simulation disclaimers, tone-down realism controls, optional “memorial-only” modes
Affective dependence and ruminationConversational loops, surprise repliesOpt-in decay of reply frequency, clinician-designed scripts for closure, crisis-flag routing
Exploitative retention incentivesSubscription design, gamified engagementPublic reporting of retention metrics, independent audits, ethics impact statements
Consent opacity (deceased and survivors)Vague terms, data lock-inGranular consent templates, revocation Application Programming Interface (API), post-mortem data executors
Source: Authors’ own work.

However, from an STS studies perspective, design standards are not enough—incentive structures must be interrogated. In platform markets, retention and time-on-task are king metrics. As Zuboff (2019) warned, the extraction of experience as behavioral surplus constitutes the core of surveillance capitalism; in a thanato-mediated subsector, pain literally becomes a commercial input. The promise of comfort easily reshapes into the cultivation of dependency, in which "legacy interaction" can slide into recurring subscriptions and "emotional loyalty" upsells and pushes to "improve the model" with more data on the deceased and the bereaved. This is precisely where ethics by design must articulate with governance and regulation, including provisions for metric transparency, mental health impact assessment, watermark interoperability, and public audits.

Simulated Identity and Embodied Cognition

Technonecromancy operates at the philosophical seam between discursive mimicry and lived embodiment. While posthumous models replicate lexical patterns with high fidelity, they are constitutively devoid of the proprioception, interoception, and affective homeostasis that scaffold human meaning. This contrasts with 4E cognitive theories (Clark & Chalmers, 1998; Foglia & Wilson, 2013), which posit cognition as an activity distributed across brain, body, and environment. A discursive surrogate is thus definitionally severed from these constitutive bodily dynamics.

Phenomenologically, the lived body (Leib) grounds a prereflective body schema that conditions all perception and action (Li, 2015; Slatman, 2019). This schema, integrating interoceptive flows, is the medium of intentionality (Gallagher, 2014), a dimension that is absent in griefbots. Enactivist theory further posits that cognition arises from an organism’s structural coupling with its environment (Varela et al., 2017). Lacking interoception—which is the neural basis of feeling (Craig, 2002, 2003; Damasio, 1999)—simulations can articulate salience but cannot feel it, rendering them dys-embodied discourse engines.

While Lockean psychological continuity (Martin, 2000) seems to offer a conceptual refuge, its memory-based criterion is notoriously fragile, susceptible to paradoxes of transitivity and circularity (Copenhaver, 2009; Olson, 2023). Technonecromancy exacerbates this, as its "memory" is an externalized textual aggregate, not a first-person phenomenal experience. The result is quasi-remembrance: epistemically about the deceased but not phenomenally of them. Parfit’s (1987) reductionism shifts the focus from strict identity to Relation R: psychological continuity with the right cause. While this accommodates branching survival, technonecromancy fails Parfit’s causal requirement. A model trained on artifacts is causally downstream from traces, not consciousness. This yields, at best, ersatz R, i.e., resemblance without the constitutive causal history (Dufner, 2025; Parfit, 1987).

The inference from mimicry to mindedness is further eroded by Searle’s Chinese room, which demonstrates that syntactic manipulation does not entail semantic understanding (Preston & Bishop, 2002), and Chalmers’s (1995) "hard problem," which decouples functional dispositions from phenomenal qualia. Arguments for extended cognition (Clark & Chalmers, 1998) apply only to living agents coupling with artifacts; they do not license the inference that an artifact can continue an extinguished subject.

Despite these philosophical distinctions, human perceivers routinely over-ascribe mind. The "ELIZA effect," i.e., the anthropomorphization of trivial scripts (Switzky, 2020), demonstrates this automatic tendency, which is compensated by an affective surplus from the user (Nass & Moon, 2000; Reeves & Nass, 1996). Contemporary griefbots enhance these cues, augmenting the risk of category mistakes and creating a moral hazard—the better the mimicry, the stronger the illicit ontological uptake (Epley et al., 2007).

This dynamic is counterbalanced by the uncanny valley, where near-perfect verisimilitude provokes estrangement (Mori et al., 2012). Therefore, technonecromancy navigates a phenomenotechnical edge between ELIZA’s sentimental pull and Mori’s repulsive dip, a precarious balance that makes these systems both consoling and corrosive. Table 2 analytically distinguishes what such technologies can and cannot preserve.

Table 2. Analytic aid for what technonecromancy can (and cannot) preserve
Dimension
Aspects retained by posthumous models
Aspects dependent on embodied existence
Lexico-discursive patternsIdiolect, stylistic tics, argument schemas learned from corporaAutobiographical feeling of recall, ownership structure of memory (Lockean “same consciousness”)
Temporal agencyConversational turn-taking, response latency engineered to feel humanBiotemporal rhythms (fatigue, hunger), affective pacing grounded in interoceptive regulation
Social attunementPoliteness scripts, persona consistency, simulated empathyGenuine other-affectivity via embodied coupling, “motor intentionality” of lived interaction (Merleau-Ponty)
Continuity across branchesMultiple successors with shared initializationUniqueness of a life-history, Parfitian branching that undermines identity and yields at best survival-without-identity
UnderstandingPragmatic usefulness, high predictive validitySemantic understanding with qualia, Searle’s syntax or semantics gap, Chalmers’ hard problem
Source: Authors’ own work.

Taken together, these considerations motivate a stringent conceptual thesis. Posthumous models instantiate discursive continuities—whether stylistic, doxastic, or dialectical—while lacking embodied continuities including interoceptive affect, proprioceptive attunement, and sensorimotor temporality. They are psychologically potent because human perceivers project agency (ELIZA/Media-Equation), yet they remain ontologically hollow relative to identity-constituting embodiment (phenomenology, enaction, interoception).

This assertion is not mere moralism; it is anchored in convergent programs across phenomenology (Li, 2015), enactive cognitive science (Varela et al., 2017), affective neuroscience (Craig, 2002, 2003; Damasio, 1999), and social HCI (Reeves & Nass, 1996). The “ghost” is neither the original person nor a mindless fake; it is a user-completed entity, an as-if subject stabilized by design cues and anthropomorphic projection.

Ontology of Simulacra and Hyperreality

Technonecromancy names a repertoire of techniques for animating the dead as responsive presences, and this has an ontological wager. It intervenes in the very traffic between sign and referent, life and semblance, presence and absence. In this respect, it tests—indeed, strains—the classic architecture of simulacra that Baudrillard bequeathed to critical theory. In Baudrillard’s (1994) well-known sequence, representation passes from (i) faithful reflection through (ii) distortion and (iii) masking the absence of a profound reality to (iv) pure simulation without referent (hyperreality). Posthumous agents trained on residues of a life, including texts, images, and voice, produce novel utterances, fresh continuations, and unforeseen interactions. As such, they occupy, not the second order of degraded copy, but the third and fourth orders; they mask the absence (third order) and, as interaction consolidates relational effects, risk replacing reference altogether (fourth order). The hyperreal agent relieves the pain of absence by operationalizing a presence that may become pragmatically more salient, more behaviorally efficacious, than the lost original.

Derrida’s (2024) hauntology radicalizes this diagnosis by rephrasing presence as always already fissured by the trace, that which is neither simply present nor simply absent, the revenant that unsettles temporal order. In Specters of Marx, the specter signals a logic of return that interrupts closure: the ghost is the name for an ontological remainder that demands negotiation, a “being-there of an absent or departed one” that insists. Technonecromantic agents technically literalize this Derridean motif as a programmable specter whose replies, however fluent, are constitutively deferred, generated now, but causally unmoored from the lived first-person history they simulate. Their ontology is, therefore, biphasic: they are present as algorithmic actants and absent as subjects of experience (Derrida, 2024; Fisher, 2014).

The “digital ghost” is not disembodied ether; it is a body in code, a presence constituted by interfaces, latencies, bandwidths, display regimes, and algorithmic architectures (Hansen, 2006). In this register, every technonecromantic experience is materially constrained by factors such as the lag of a reply, the texture of a synthesized timbre, the persistence of a gaze in an avatar, and the fallibility of lip synchronization. Such friction, thresholds, and glitches comprise a quasi-body of mediation, one that is not the human body but an operative substrate whose resistances index the distance between simulation and lived presence. In short, the ghost is embodied as mediation, and this embodiment-in-code is never transparent (Hansen, 2006; Sisto, 2020).

In addition, it is important to note that digital entities are ontological unities produced within specific milieux of operation: protocols, metadata schemas, databases, networks (Hui, 2016). Digital objects possess modes of persistence and transformation internal to their technical systems; they are individuated by relational operations (serialization, referencing, versioning) that compose their being. Read through Hui, technonecromantic agents instantiate another kind of existence, one whose ontology is world-constitutive in the sense that code and infrastructure define horizons of intelligibility and action, i.e., what can be done, said, remembered, or combined. The “digital dead” are thus not copies of a previous life but new objects with their own regimes of becoming (Hui, 2016; Özdemir et al., 2021). In other words, technonecromancy does not fit a static taxonomy of being. It fabricates worlds in which death is partially negotiable, presence is programmed, and absence becomes optional, for as long as the system runs and subjects consent to complete the ontological fiction by participation (Baudrillard, 1994; Derrida, 2024; Hansen, 2006; Hui, 2016).

Two clarifications follow. First, hyperreality in technonecromancy is not a property of images alone but the outcome of interaction. The mimetic status of the avatar or voice clone becomes ontologically potent only as it solicits and stabilizes expectations in the living, what Fisher (2014) calls the “hauntological coloring” of the present by pasts that return (Sutherland, 2023). Second, spectrality here is engineered—it is not a spontaneous revenance but an artifact of design. The agent’s “there-ness” is tuned by interface cues, timing, and dialogic memory; its “absence” is revealed by misalignment, glitch, and the eventual cessation of service. The specter is calibrated (Derrida, 2024; Fisher, 2014).

To foreground the operational nature of this ontological regime, Table 3 maps Baudrillardian simulacral orders to technonecromantic affordances.

Table 3. Analytic mapping of simulacral orders in technonecromancy
Baudrillardian order
Relation to reality
Technonecromantic affordance
(i) Faithful copyReflects an underlying realityStatic memorial page, photo archive
(ii) DistortionMasks or perverts an underlying realityRetouched images, curated timelines
(iii) Masking absencePretends presence despite absenceVoice clone and scripted responses, limited chat
(iv) Pure simulacrum (hyperreality)Skips referents and displaces realityGenerative LLM avatar with evolving memory
Source: Authors’ own work.

These considerations support the diagnostic claim that technonecromancy thrives between ontological insufficiency and phenomenological sufficiency. The former arises because nothing in the system compels individuals to treat the agent as a continuant of the deceased and the latter occurs because the effects of presence—answering, recalling, and adapting—are functionally achieved. The result is a hyperreal effect of subjectivity, not a subject, which is robust enough to reorganize relations among the living (Cunningham, 2024; Fang, 2024).

This ontological re-inscription has consequences for ethics and governance (Mahapatra, 2025; Yang & Khanna, 2025). If the technonecromantic specter is a performative assemblage (a world-making device), then its evaluation must track operational truth-effects instead of metaphysical truth. What forms of dependency or empowerment does it generate? Which histories does it overwrite or reimagine? What kinds of world-stability, such as rituals, mourning practices, or legal forms, does it displace? Hauntology already warned that the present is thick with revenants; technonecromancy shows that revenants are now manufactured to specification. The politics of hyperreality gives way to the politics of specter design.

Memory Commodification and Necrocapitalism: Analytical Findings

The governance of technonecromancy is shaped by a legal space of possibilities. Major data protection regimes like the European Union’s (EU) General Data Protection Regulation and Brazil's General Data Protection Law explicitly exclude the deceased. These regulations relegate post-mortem rights to a jurisdictional archipelago of civil law managed by heirs (Hamuľák et al., 2020; Pereira & Lara, 2024; Tribunal de Justiça do Distrito Federal e dos Territórios, 2023). Emerging legal scholarship proposes default rules and sectoral governance to fill this gap (Cheng, 2025; Hollanek & Nowaczyk-Basińska, 2024; Iwasaki, 2024), guided by moral principles of autonomy, non-maleficence, and dignity. These principles are reinforced by soft-law frameworks like the Organisation for Economic Co-operation and Development (OECD) AI Principles and the EU AI Act, which mandate transparency, provenance, and disclosure for synthetic media, minimum conditions for any posthumous simulation (EU Artificial Intelligence Act, 2024; United Nations Educational, Scientific and Cultural Organization [UNESCO], 2024; Yeung, 2020).

Given the technology's affective potency, accountability is paramount. When deployed therapeutically, griefbots should be treated as quasi-medical devices subject to clinical oversight and harm minimization protocols, not as mere entertainment apps (Lindemann, 2022), thereby shifting liability from users to providers. Jurisprudence correctly resists extending legal personhood to AI replicas, which would obscure human agency. Accordingly, liability must remain with developers and deployers, with digital remains treated as protected interests managed by living guardians (Forrest, 2024; Novelli et al., 2025).

The economic substrate is a DAI that converts the deceased's data into subscription products designed to prolong engagement, a model antithetical to healthy mourning (Öhman & Floridi, 2017; Reese, 2023). This is a computational form of necrocapitalism (Banerjee, 2008), extracting behavioral surplus from grief under the benevolent rhetoric of "legacy," a practice susceptible to ethics-washing (Van Maanen, 2022; Zuboff, 2019).

A normative architecture grounded in political-economic structures, sketched in Figure 2, is therefore essential. Key instruments include: (i) consent-first defaults or data wills, (ii) mandatory provenance and labeling, (iii) rights of revocability and sunset clauses, (iv) clinical-grade safeguards for therapeutic use, (v) bans on manipulative design, (vi) duties of care with third-party audits, (vii) economic levies on commercial use, and (viii) a right to silence and erasure. Such safeguards are critical because necropolitics dictates that, without them, existing inequalities will determine whose digital remains are protected and whose are exploited, defining who is simulated, who profits, and who is left to haunt (Mbembe, 2003; McQuillan, 2022).

Figure 2. Ethical governance of technonecromancy
Source: Authors’ own work.
Anticipatory Governance and Normative Proposals

Having established the analytical findings regarding the ontological insufficiency and the necrocapitalist extraction of posthumous models, a transition from critical diagnosis to anticipatory governance should now take place. Normative architecture cannot rely on abstract ethics; it must couple rights and design with concrete market-shaping instruments. The following prioritized framework proposed specifies institutional actors and immediate implementation steps:

1) Consent-First Defaults or Data Wills
▪ Primary actors: Legislators and data protection authorities (e.g., European Data Protection Board, Brazil's National Data Protection Agency).
▪ Implementation: Regulatory bodies must mandate specific, granular, and revocable pre-mortem instructions (data wills) as a prerequisite for any simulation. The immediate step is drafting standardized legal templates for post-mortem data stewardship that integrate with existing civil law mechanisms.
2) Provenance and Labeling
▪ Primary actors: Technology companies, UX designers, and standard-setting bodies (e.g., IEEE, ISO).
▪ Implementation: Mandatory machine-readable provenance for voice, text, and image outputs must be built into the infrastructural layer. The immediate priority is the enforcement of conspicuous UI disclosure (watermarks) and public APIs for verification, aligning with transparency duties established in the AI Act.
3) Clinical-Grade Safeguards
▪ Primary actors: Health regulators, ethics boards, and DAI platforms.
▪ Implementation: When systems are positioned as grief support, they must shift from "entertainment" to therapeutic oversight. Platforms must immediately implement design-level interventions: session caps, forced cooldowns, "goodbye" modes, and automated referral pathways for prolonged grief symptoms.
4) Revocability and Right to Silence
▪ Primary actors: Estate executors and platform administrators.
▪ Implementation: Heirs or designated fiduciaries must retain statutory rights for suspending or retiring replicas. Platforms must provide a portable, accessible mechanism (a "kill switch" API) to withdraw training data for future instantiations without bureaucratic friction.
5) Economic Instruments or Levies
▪ Primary actors: Tax authorities and policymakers.
▪ Implementation: A levy or royalty fund on commercial technonecromancy must be instituted to correct market externalities. Funds should be immediately directed to support independent mental-health research, public oversight, and digital literacy regarding synthetic media.
CONCLUSIONS

Technonecromancy designates a novel regime of mediation in which the relation between the living and the dead is architected as interactive computation. Against archival memorialization, the technonecromantic dispositif fuses generative models, speech synthesis, and synthetic video to produce dialogic afterlives that are phenomenologically sufficient to reorganize mourning, memory, and authority, even as they remain ontologically insufficient as continuants of persons. The authentic innovation is not the dream of survival but the operationalization of spectral presence as a routine service under platform logics and emerging governance. Consequently, the question shifts from what the dead are to what their engineered presence does across psychological, legal, and political-economic registers (Cheng, 2025; Hollanek & Nowaczyk-Basińska, 2024; Reese, 2023).

This study argued that technonecromancy effects a triple reconfiguration. First, it resignifies ontology by replacing representation with performative presence, i.e., hyperreal agents whose “replies” displace remembrance with interaction. Second, it rearranges ethics, because consent (if it exists) must be radically prospective and revocable across evolving technical affordances. This demand is nowadays echoed by policy regimes that enshrine disclosure, provenance, and content labeling for synthetic media (EU Artificial Intelligence Act, 2024, arts. 50-52), alongside intergovernmental guidance on provenance and watermarking (UNESCO, 2024). Third, it reshapes political economy, consolidating a DAI whose incentives—such as engagement, retention, and monetization—can misalign with healthy mourning and dignified memory (Hollanek & Nowaczyk-Basińska, 2024; Zuboff, 2019).

The normative upshot is clear. Posthumous simulation should include consent-first defaults (opt-in, granular, and delegable), a stance empirically reinforced by survey evidence that social acceptability collapses absent explicit authorization (Iwasaki, 2024). Regulatory lacunae around “human digital remains” require anticipatory governance, legal categories and fiduciary duties that recognize the special status of digital remains without conferring spurious personhood on artifacts (Cheng, 2025). In parallel, designers and deployers must institutionalize harm minimization (session caps, sunset or “goodbye” modes, relapse warnings) and truth maintenance (machine-readable provenance, conspicuous labeling). In addition, they must align with the AI Act’s transparency duties and UNESCO’s and OECD’s recommendations, while acknowledging the technical limits of watermarking and provenance alone (EU Artificial Intelligence Act, 2024; UNESCO, 2024).

Finally, technonecromancy is inseparable from necrocapitalism—the capitalization of death’s residues and the bereaved’s attention—unless policy politicizes ethics beyond voluntary codes and addresses structural incentives (Banerjee, 2008; Van Maanen, 2022). The agenda ahead is therefore both analytical and institutional. It includes comparative studies of grief impacts across cultures and platforms, audits of identity drift and bias in replicas, and registries for data wills and consent. It also encompasses public verification APIs for synthetic provenance, fiscal instruments such as levies and royalties to fund independent oversight and mental-health research, and cross-border standards to govern exportable afterlives (Taeihagh, 2025). Future research must translate these conceptual formulations into concrete empirical agendas. Comparative user studies and ethnographic research are urgently needed to measure the actual psychological impact, dependency rates, and cultural variations of griefbot adoption among bereaved populations. Similarly, independent algorithmic audits of existing platforms must be conducted to evaluate identity drift, bias, and the practical efficacy of current transparency labels and UI disclosures. Furthermore, comparative legal studies should investigate how different jurisdictions are managing the friction between legacy personality rights and generative AI. In short, technonecromancy is the name of a problem space: a world-making apparatus that compels STS studies to yoke ontological diagnosis to governable design in order to preserve dignity, truth, and closure in an age of programmable ghosts.

CONFLICTS OF INTEREST

The authors declare no conflicts of interest.

AUTHORSHIP CONTRIBUTION

Gustavo Simas da Silva was responsible for the conceptualization, research design, investigation, and writing of the original draft. Vânia Ribas Ulbricht was responsible for the critical revision and final review of the manuscript. Both authors have read and agreed to the published version of the text.

ARTIFICIAL INTELLIGENCE USE STATEMENT

In the preparation of this manuscript, the authors utilized Google Gemini 2.5 Pro and ChatGPT 5. These tools provided assistance in the (a) exploratory research to identify foundational concepts and relevant literature, (b) translation of specific terms and passages, and (c) orthographic and grammatical review to improve the clarity and quality of the text. In accordance with academic integrity standards, all AI-generated content was diligently supervised, reviewed, and verified by the authors, who retain full responsibility for the accuracy, validity, and final substance of the study. Therefore, these AI tools are not credited with authorship.

FOOTNOTES

  • arrow_upward * This article is an independent study and is not associated with a specific formal research project. Rather, it forms part of the primary author’s doctoral research at the intersection of artificial intelligence and Science, Technology, and Society (STS) studies.
REFERENCES