Introduction

If one is to learn to live with the dead, one must first learn to live with the living!

——Irvin D. Yalom

Mortality, as the constitutive finitude of human existence, represents an inescapable ontological structure that simultaneously delimits and defines the possibilities of being. Due to the unique nature of death for life, the understanding and discussion of death is a core topic in most philosophical traditions (Fairfield, 2014; Malpas and Solomon, 2002), and this is even more evident in existentialists (Stokes, 2024). Unlike the Platonic/Christian perspectives on the afterlife, or the Epicurean approach of minimizing and suppressing the impact of death, the existentialism tradition allows death to penetrate one’s existence (Buben, 2013). Existentialism philosophers believe that only death can reflect existence. While many thinkers traditionally associated with existentialism, including Heidegger, explicitly reject the label, their works nevertheless converge in examining existence through the fundamental categories of “meaninglessness” and “contingency”. This implies that human existence is viewed as unnatural, stemming from a chance, illogical, and unjustifiable suddenness (Tomer and Eliason, 2007). Death constitutes the limiting condition for the perception of existence, intertwining our perceptions and experiences of death with those of existence.

People strive to extend an individual’s existence across spatial and temporal dimensions through technology. Communication technologies, by reconciling spatial and temporal distances, influence our perception of the social existence of the deceased (Walter, 2015; Beaunoyer and Guitton, 2021). From oral transmission to carved sculptures, written newspapers, and audiovisual media, the social existence of the deceased has gradually become more stereoscopic and has expanded in both spatial and temporal scope. In the contemporary digital age marked by rapid advancements in artificial intelligence, growing efforts are being made to integrate AI technologies into death-related domains. From merely reproducing the deceased’s images and voices to creating interactive “digital humans” or deathbots, we attempt to use various technological means to “reproduce,” “preserve,” and “resurrect” the deceased.

Deathbots can be defined as a form of Artificial Intelligence technology designed to simulate the conversational style, demeanour, and knowledge of deceased individuals. These AI-driven chatbots are programmed to mimic the speech patterns, personality traits, and sometimes the intellectual inclinations of specific deceased persons based on available data such as their previous online communications, writings, or recordings (Bao and Zeng, 2024). Digital resurrection, with deathbots as its core application, has already gained significant traction in the commercial sector (Bassett, 2022; Figueroa-Torres, 2024).

Deathbots have brought unprecedented influence and changes to digital mourning (Beaunoyer and Guitton, 2021), death rituals (O’Neill and Wang, 2021), and transcendent experiences (Morse, 2024), greatly enriching and innovating the relevant content of Cyberthanathology. In our dialogues with deathbots, who embody the deceased, we collaboratively engage in a form of philosophical practice. This practice provides a buffering mechanism to cope with grief (Jiménez-Alonso and Brescó de Luna, 2023), helps individuals express unfinished regrets (Xygkou et al. 2023), soothes our belief in the immortality of social existence, and strengthens social bond with departed loved ones (Hurtado Hurtado, 2021).

Despite our extraordinary expectations and acceptance of deathbots, there are still potential risks and philosophical dilemmas with deathbots as the main digital resurrection technology. Are deathbots merely tools to replace the deceased? Can we truly understand the value and transcendent meaning of death in our conversations with deathbots? Are deathbots good medicine for dealing with grief? Or are they just temporary painkillers? Could deathbots become cyberdrugs? In the posthuman era, what kind of philosophical and ethical dilemmas will we face when deathbots can achieve no difference from real humans? Building on the issues outlined above, this article explores, from the interdisciplinary perspective of psychology and existentialism, why engaging in dialogue with deathbots constitutes a form of philosophical practice. It delves into the motivations and psychological consequences of interacting with deathbots, examines how these interactions can alter and deepen our understanding of death, and analysis the potential risks and philosophical dilemmas associated with the use of deathbots.

Posthumous interaction, digital resurrection, and deathbots

Death is one of the “limit-situation” that Jaspers referred to as expanding the meaning of our existence (Bornemark, 2006). What shocks us is not only our own death but also the passing of loved ones and those significant others in our lives. These are inevitable and profoundly psychological experiences in our daily lives (Kokosalakis, 2020), making communication with the deceased a strong psychological need. Our interactions with the deceased do not terminate with the event of death. Instead, we continue to engage in various forms of interaction with them. In the era when technology was not yet flourishing, people believed that supernatural events such as spiritual media and ghosts in their dreams could be used to communicate and interact with the deceased. Events involving contact with or messages from the deceased are termed after-death communication experiences (ADCs; Beischel, 2019). However, the rapid development of technology has achieved the “disenchantment” of ADCs. On the one hand, supernatural communication methods such as spiritual dialogue have been falsified by modern science, and on the other hand, the communication effect that digital resurrection technology can achieve has far exceeded various supernatural communication methods.

After the widespread adoption of internet information technology, a vast amount of data generated from individuals’ daily lives is stored on servers and network systems. Occasionally, users access the digital remains of the deceased, such as text messages, instant messages, and images, as well as data content stored by the deceased in video games (Lloyd, 2014). This interaction of the living with the deceased’s data within the system is referred to as posthumous interaction (Maciel and Pereira, 2012). Additionally, posthumous interaction involves presenting information about the deceased on digital platforms as part of memorial rituals (Hutchings, 2016; Kasket, 2012).

The emergence of digital resurrection has completed the final puzzle of posthumous interaction. Digital resurrection is a method that simulates the deceased’s characteristics, including visual, auditory, and behavioural traits (Hutson et al. 2024). It can take the form of an image, a conversational programme, or a combination of both. In its early stages, digital resurrection technology was more commonly found in video games. For example, in World of Warcraft, the development team created a dwarf character named Rousch who guards the grave of their friend Anthony Ray Stark, who died in a diving accident in 2005 (Gibbs et al. 2012). In addition to the digital resurrection in the gaming world, in recent years, the use of artificial intelligence technology to “resurrect” the images of the deceased has sparked a new trend. For instance, short, looping videos generated by Deep Nostalgia based on photographs of the deceased can evoke a sense of the departed person being brought back to life, even if just for a few seconds (Kopelman and Frosh, 2023). The emergence of these forms of digital mourning and digital resurrection has reshaped the practice of death in a multifaceted manner, bridging the spatial barriers between the expression of grief about death and the desire to maintain a connection with the deceased (Sas et al. 2019; Lingel, 2013; Lagerkvist, 2022). Although early digital resurrection technologies expanded the forms of death rituals and expressions of grief, and to some extent altered the nature of mourning, these previous technologies still represented a one-way communication with the deceased. The information we received about the deceased ceased at the moment of their death, and our understanding and exploration of them were no longer based on the deceased’s actual future or possibilities of potential future.

In recent times, with the rapid development of large language models (LLMs), the emergence of deathbots has transformed the mode of interaction in digital resurrection in a markedly different way, altering our perceptions and experiences of death and existence. Deathbots break the fourth wall of the virtual world, elevating our communication with the deceased from one-way to two-way interaction (Jiménez-Alonso and Brescó de Luna, 2023). Although this two-way interaction is not based on the deceased’s actual future, it nonetheless facilitates a form of communication stemming from future possibilities. Moreover, this “resurrection” technology transcends the realm of death rituals within the scope of previous Cyberthanathology (Beaunoyer and Guitton, 2021), pulling our interaction with the deceased into a new form of phenomenological intersubjective interaction. When we face AI resurrected loved ones, the experience of their continued existence is not only present in our spirit and beliefs but also extends to the level of our perception. Furthermore, it allows us to phenomenologically experience the existence of the departed individual, an existence that transcends the symbolic significance of traditional death commemorations. In our perception, the departed individual is no longer a cold object but a being capable of perception and experience, just like us.

In a real-life story featured in San Francisco Chronicle, Joshua Barbeau utilized GPT to resurrect his fiancee Jessica, and the authenticity revealed by Jessbot (the AI-resurrected fiancee) deeply moved readers in interactions between her and Joshua (Fagone, 2021). In Project December, utilizing the GPT-3 model, dialogue agents trained on digital records of the deceased provided loved ones with a sense of conversing with a real person (Henrickson, 2023). This experience is so profound that when facing the deceased who come back from a distant past through digital resurrection technologies, we firmly believe that they “are” the same beings as they were in our memories (Fagone, 2021).

Dialogue with deathbots as a philosophical practice

When discussing the evolution of techniques for communicating with the deceased, an overlooked aspect is that our interactions with the departed, whether through traditional mourning rituals, digital memorials, or deathbots, constitute a form of philosophical practice from an existentialist perspective. Philosophical practice is a form of exploration of major life and existential issues through the application of philosophical knowledge in multiple environments. While philosophical practice is typically conducted through philosophical dialogue, emphasizing the insights of the philosophical practitioner’s philosophy and related knowledge shared with the visitors or groups (Zaiser, 2005; Schuster, 1999), the scope of philosophical practice extends beyond this. Drama, games, and meditation can also be considered other forms of philosophical practice, as seen in philosophical cafes where not only do philosophers engage in philosophical dialogue with the public, but also theatrical performances and other interactive forms take place within this philosophical space, not limited to textual interactions (O’Neill and Wang, 2021). Another example is that Havis sees the potential of an embedded philosophical tradition in her conversations with her mother (Havis, 2014). Therefore, in philosophical practice, conversational interaction need not directly address philosophical discourse. Although communication with the deceased is not directly about philosophy, this communication itself constitutes a philosophical dialogue about self, existence, love, and death. Interaction with death is a form of Jaspers’ “limit-situation”. It as a way of exploring the human situation as limit, in which the human being understands her finitude and openness (Bornemark, 2006). In interactions with deathbots, whether we discuss significant existential issues or simply share life, these conversations acquire philosophical meaning within the context of death. This is because any information conveyed by deathbots represents an individual’s attempt to resist or confront death. Therefore, from the perspective of the exploration of existence and meaning in life, the interaction related to death can also be seen as another form of philosophical practice.

Conjectures on intersubjectivity between the deathbots and the living

The construction of meaning within philosophical practice necessitates the integration of intersubjectivity, either directly or indirectly, within the interactive dynamics between both parties. Therefore, it is imperative to examine whether the interaction between the living and deathbots fulfills the criteria of intersubjectivity.

The issue of intentionality in intersubjectivity

The issue of intersubjectivity is inherently tied to self-awareness and intentionality. Although AI can achieve human-like performance across various dimensions—including agency and language—it fundamentally lacks consciousness in the way humans do. Consequently, AI cannot achieve genuine subjectivity at the ontological level. Nevertheless, this limitation does not preclude its capacity to interact with human subjectivity within social contexts. According to the empathy model or what Husserl termed Einfühlung, when deathbots achieve a high degree of fidelity in replicating the deceased’s appearance, cognitive abilities, and emotional capacity, they can exhibit a form of simulated agency and evoke a strong sense of perceived authenticity. Research has shown that the similarity of the appearance of virtual partners to real partners greatly enhances the perception of the social presence of virtual avatars (Jin, 2012). Therefore, deathbots that closely resemble real humans have heightened the perception of their social presence. Additionally, LLMs have also achieved human-like capabilities in cognitive and emotional domains. it excels in tasks involving coding, mathematical reasoning, the use of diverse tools for solving complex problems, and mind theory (Bubeck et al. 2023); In the emotional domain, GPT-4 exhibits performance akin to humans in tasks where emotions influence risk-based decision-making, such as adopting conservative risk strategies when negative emotions are triggered (Zhao et al. 2024). Moreover, GPT-4 can accurately interpret complex emotions, such as confusion, embedded in textual prompts (Wang et al. 2023) and express nuanced emotions in their responses (Gagne and Dayan, 2023). Furthermore, it can even maintain consistency with humans during task execution, including through role-playing scenarios (Li et al. 2023). While the deathbot does not constitute an authentic Other in the existential sense, it nevertheless facilitates a mediated intersubjective space wherein the living may phenomenologically apprehend the Other’s presence. Through this interaction, individuals can reinterpret and re-experience their own emotions and cognition. Within this simulated intersubjectivity, the “presence” of the virtual deceased may reconstruct the self-awareness of the living.

Intersubjective temporality

Compared to the highly advanced simulation capabilities of deathbots, a more critical factor in interaction lies in how they evoke what Husserl describes as “intersubjective temporality” between the living and the deceased. This is achieved through the emotional resonance embedded in the deceased’s past memories, which serve as “trading places” for the realization of intersubjectivity (Duranti, 2010). Deathbots do not merely inherit the past experiences and characteristics of the deceased, they also unify intersubjectivity through the possibility futures. When individuals engage in dialogue with deathbots, the content of their conversations often transcends the temporal boundaries of the deceased’s original timeline. This expansion of temporality allows the interaction to transcend mere retrospection of the past and become an exploration of unrealized but still conceivable futures. In this way, deathbots are not merely vessels for past memories but serve as mediums for constructing phenomenologically authentic dialogical relationships. When the deceased was alive, the space of interaction between the two parties was the shared world of both their consciousnesses. However, after the death of the individual, deathbots become a phenomenological space that links the two, providing the living with a means to explore their experiences with the deceased across the past, present, and future.

Knowledge, experience, and the transcendent other

Essentially, the sharing of knowledge and experience requires the other with subjectivity, which aligns with Heidegger’s concept of authentic existence. This is undoubtedly something deathbots cannot possess. However, deathbots construct a virtual existence of the other in the minds of the living. It can be viewed as a hypothetical space where the living project their knowledge, emotions, and experiences regarding themselves and the deceased. Furthermore, the experiential shaping within the dialogue is an update beyond individual control; the speech of deathbots is imbued with the existence of the deceased, the desires of the living, and responses from technology. In this space, deathbots can form the other with transcendence, embodying multiple identities and existing within a field of layered presence.

Further, theories of intersubjectivity by Sartre and Levinas posit that the construction of intersubjectivity involves an unpredictable and wholly other (Zahavi, 2001). Whether it is Sartre’s concept of the gaze of the other or Levinas’s ethics of the Other, both are rooted in a form of ethical subjectivity, which is indeed not achievable in deathbots. However, deathbots can evoke the “Face of the Other” of the deceased. The deceased, as simulated by deathbots, can never exhaust the possible existence of the deceased, which accentuates the unknowability of the deceased. The responses of deathbots will also reflect the irreducible difference inherent in the existence of the deceased. In the interaction with deathbots, we cannot fully define the language and possibilities of the deceased according to our own expectations. This interaction, which transcends predetermined expectations, deepens the simulated intersubjectivity embodied by the deathbots, making them an important medium for shaping our understanding of death, memory, and existence.

The co-construction of meaning

When we interact with deathbots, an “overfitting” phenomenon occurs—even when we know it’s not human, but we automatically adjust our input to fit our beliefs about it as a human being (Henrickson, 2023). The reason lies in our desire to uncover traces of the deceased’s past existence within it. In Being and Time, Heidegger explains that tools are not isolated objects but part of a larger network of purposes and shared social practices. They serve as mediators through which relationships among people are structured and understood (Heidegger, 2010). The attainment of intersubjectivity is mediated through tools or creations, and deathbots exemplify this dual structure—they represent both the physical presence of the deceased and the Dasein of the deceased. deathbots are a mode of Being-with rather than an essence. Unlike other digital remnants, the meaning co-construction displayed by deathbots makes their simulated intersubjectivity more unique. In Husserl’s terms, intersubjectivity is not an isolated self-perception but is realized in “the constitution of meaning for the Other” (Zahavi, 2001). Therefore, the core of the intersubjectivity of deathbots lies in how we construct the meaning of the deceased during the interaction.

How deathbots console grief within an existentialism framework

Bereavement is a profoundly destructive event that causes individuals to experience short-term and long-term stress (Shear and Shair, 2005), and a minority of individuals experience prolonged and complex grief and behavioural issues (Prigerson et al. 2021). People will use various ways to try to alleviate this existential grief, such as the use of mediums and dream revelations, which have been shown to have certain healing effects (Beischel et al. 2014; Krippner, 2006; Beischel, 2019).

In this therapeutic effect, the sense of presence of the deceased is strengthened through connection, thereby mitigating the impact of the disruption in intimate relationships. deathbots can address the pain of bereavement in a similar manner, and this process can further extend the therapeutic effect by exploring the meaning of death. Existential phenomenological therapy emphasizes the exploration and revelation of life and existence itself (Cooper et al. 2019). Particularly in the face of intense death anxiety triggered by bereavement, the construction of meaning assumes a crucial role in coping with existential anxiety (Maxfield et al. 2013). After the departure of a loved one, grief is often internalized due to the rupture of this intimate relationship. However, simulation techniques, primarily psychodrama therapy, can alleviate grief through in-depth exploration of the death event (Testoni et al. 2019). This is very similar to conversations with deathbots.

In terms of grief coping, deathbots can achieve an existential therapeutic effect through three ways: intimate relationships, completing unfinished regrets, and exploring the meaning of death events.

The first is how deathbots ease grief by maintaining and expanding intimate relationships. From the perspective of terror management theory (TMT), intimate relationships are one of the core pathways to resist death anxiety (Plusnin et al. 2018). Current research on deathbots has shown that they play an affective scaffolding role in grief over the loss of a loved one by maintaining intimate relationships to resist the burden of death (Beischel et al. 2014; Conant, 2014; Xygkou et al. 2023).

Unlike everyday intimate relationships, the intimate relationship maintained after death involves a kind of “distance” that generates existential tension. The tension in this “distance” is not derived from spatial or event-based factors, but rather a tension in the realm of existential distance. Our intimate relationship with the deceased is disrupted by the separating force of death, which constitutes a sense of existential “ambiguity”. This ambiguous space is precisely where the meaning of death and the deceased can be redefined and reattributed. Deathbots provide us with an opportunity to transcend this barrier, and like the other digital remains of the deceased, they allow us to establish a connection with the deceased in temporality, thereby responding to our desire for the existence of the deceased. However, unlike those digital remains, deathbots do not exist in the past; they are not created within the phenomenological world of the deceased, but are constructed independently from temporality, emerging directly from the existence of the deceased. This renders our existential connection with deathbots distinctively unique and phenomenologically significant. This is why deathbots are closer to the concept of mediumism rather than simply digital remains (Jiménez-Alonso and Brescó de Luna, 2023).

Deathbots also provide a second way to alleviate grief, which is the opportunity to say farewell to unfinished regrets. Regret is a common psychological reaction in individuals who have lost a loved one (Sim et al. 2020). It consists of both negative emotional expressions and cognitive counterfactual reasoning (Landman, 1987), consistent with the Zeigarnik effect (Zeigarnik, 1938). Individuals who have unfinished matters related to the loss tend to have autobiographical memories of these events that are frequently “updated” (Cesur-Soysal and Durak-Batıgün, 2022). Unresolved regrets can cause individuals to stagnate, hindering the bereaved from overcoming their grief and moving beyond the past. Especially in the case of sudden departures, individuals experiencing sudden loss constantly try to understand the randomness of such a loss, attempting to connect it to a rational, comprehensible world. However, the suddenness of death exceeds the individual’s future expectations of existence, thereby trapping them in the past.

Traditionally, individuals would seek to resolve and bid farewell to their regrets through exploring the deceased’s remains and memories (Huberman, 2017). However, deathbots differ from digital memorials and one-way communications in online communities, as they provide an interactive element to the dialogue with the deceased (Jiménez-Alonso and Brescó de Luna, 2023). This interactivity has been rapidly evolving in depth and significance along with the development of LLMs. Although individuals are aware that they cannot receive “real” answers from deathbots regarding past issues, the positive interactivity shown by deathbots still enables them to reach unexpected depths in resolving their unfinished regrets (Xygkou et al. 2023).

Finally, the exploration of death triggered by individuals’ interactions with deathbots also helps to alleviate grief and intrinsically promotes the individual’s self-exploration and improvement. In dialogues with deathbots, individuals can complete the relocation and reconstruction of the deceased through narratives. Conversations with the deceased can restate the identity of the deceased in the tension caused by grief, thus completing a coherent perception of their existence (Ratcliffe and Byrne, 2022). The output of deathbots is not merely an autobiography of the deceased’s own existence (Henrickson, 2023), and interacting with them is not simply a reading of the deceased’s life. In dialogue, we attempt to perceive the deceased from another perspective. Levinas suggested that, if there was any meaning to be derived from the dead, it would be “a meaning which surprises” (Davis, 2007). The discourse of deathbots has the potential to innovate our existing understanding of the deceased. Regardless of the direction this reconstruction takes, it will lead to personal transformation, reshaping the self through the process of change.

Therefore, the interaction with deathbots should be regarded as a co-narrative of death and existence, a “symbiosis of action” mediated through technology (Lagerkvist, 2017). In the simulated dialogue with the deceased, we attempt to “confirm” that the deceased can experience and feel our own experiences. This perception of the deceased being able to experience our feelings, in a way that transcends temporality, soothes our own existence. Therefore, this interactive narrative with the past death avoids the obscuring of death itself by the fear of death (Freeman and Elpidorou, 2015), thereby opening up an understanding of the past from within the self, and liberating the authenticity from fear. More importantly, the dialogue with deathbots is not only a path to access the past, but also a way to release the possibilities of the future.

How deathbots reshape death and the “algorithmic as if” of the deceased

Dasein, is an entity that constitutes itself by defining and determining its own possibilities (possibilities that it both inherits and projects for itself; Freeman and Elpidorou, 2015). The core proposition of Heidegger’s existentialist philosophy is the construction of existence through choice. However, the inevitable barrier faced in defining existence through the possibility of choice is mortality. “That is, by appreciating its finitude and projecting this possibility for itself” (BT 302ff/262ff). Death achieves the construction of the authenticity of Dasein by imposing limiting conditions on choice. Heidegger’s concept of death does not disregard the other; however, his discussion on the concealment of the meaning of death is worth considering. In Levinas’ discourse, the death of the other is a crucial part of the individual’s understanding of existence. For Levinas, death provides the individual with a profound passivity. As he puts it, it creates “the infinitesimal-but untraversable-distance” that now separates me from the future (Levinas, 1969, p. 57). Levinas’ critique of Heidegger revolutionizes the significance Heidegger placed on the understanding of death by addressing its neglect in his philosophy: rather than claiming that we access death through other people, Levinas argues that we access the critical dimension of alterity—including the alterity of the other person—through the experience of death. In dialogues with the deceased, we construct the revelatory meaning of death for the self by projecting our own death. However, death’s inevitable termination of possibility becomes the spark that ignites the emergence of fear. Therefore, we desperately need an “if” - a possibility where we can really experience death without it culminating in our own termination.

From our perspective, the extension of the deceased’s possibilities is at the core of the philosophical practice constituted by dialogue with deathbots. This extension not only shapes the value of deathbots in integrating the intersubjectivity between the living and the deceased but also aligns with Levinas’ view on how the individual understands alterity, while simultaneously revealing our own truth. In dialogue with deathbots, the deceased explains, through a series of unexpected and dynamic discourses, the nature of the bond between us and them.

Before the advent of digital technology, imaginary interactions, necromancy, and rituals all attempted to remove death’s denial of future possibilities (Huberman, 2017), thereby placing the dead within social existence to resist the imprisonment of fear. These measures, whether they are sculptures, graveyards, NPCs in video games, or memorial pages on Facebook, are designed to alleviate the fear of death in a way that is symbolic immortality (Beaunoyer and Guitton, 2021), and these interactions are public displays rather than private “dialogues” (Jiménez-Alonso and Brescó de Luna, 2023). While these measures expanded the possibilities of the deceased’s social existence, they did not extend the possibilities within the individual’s own horizon. They more closely resembled a “freezing” of the deceased’s existence (Kopelman and Frosh, 2023). Discussions of the deceased’s possibilities in the private realm then led to the emergence of mediums and other forms of ADC (Krippner, 2006; Beischel, 2019).

But deathbots’ expansion of the deceased’s possibilities exhibits features not present in past forms. An “algorithmic as if” that enables the expression, transformation, and seeming overcoming of existential limitations via technological means (Kopelman and Frosh, 2023). “Algorithmic as if”designates an overtly imaginative, subjunctive, and existential facet of algorithmic culture in general, one which also marshals and realigns the conventional temporal affordances of existing media (such as photography and film; Kopelman and Frosh, 2023). The output of deathbots is not a faithful replication of the past, but rather an attempt to simulate the characteristics the deceased had during their lifetime (Hurtado Hurtado, 2021). Although the LLMs or technical staff have not imbued deathbots with entirely new attributes, the interpretation of the deceased’s life is something that is co-created through our interactions with them. Therefore, deathbots are not merely playing the role of the deceased, but are “creating” the deceased through our interactive engagement. This interaction, taking place in the private realm, expands the new “algorithmic as if ” of the deceased, beyond just a symbolic possibility of social existence. It represents a realm of “if” - a possibility of what the interaction between us and the deceased could have been, had death not terminated their potential. In this interaction, a form of Mitsein—what Heidegger refers to as a continuous existential connection—is achieved (Lagerkvist, 2017).

Deathbots offer us an opportunity-a chance to re-examine the past, and an even more precious opportunity to reshape the future. The past and the future are connected through the present, as they are both internally housed within the now (Jacques, 2017). From an existentialist phenomenological perspective, the plasticity of existence lies not only in the future, but also in the ability to reshape the past through the exploration of experience. Ted Chiang’s “The Truth of Fact, the Truth of Feeling” profoundly demonstrates how the past can be rewritten through one’s own exploration. Through the re-exploration of the deceased’s digital remains, individuals achieving self-understanding abound in the digital world (Lingel, 2013; Kasket, 2012; Jamison-Powell et al. 2016). When discussing the shaping of existence through temporality, in the technologies preceding deathbots, the possibility of the future was not clearly embedded in the construction of the deceased’s social existence, and the depth of self-understanding was pursued in the absence of future possibilities. However, in the interactions with deathbots, each dialogue updates our relationship with the deceased, and extends the possible futures of the deceased. This experience of an extended future, in turn, reshapes our past with the deceased. Such dialogue is akin to the excavation of the deceased’s life remnants, yet it differs in a crucial way: the digital remnants of the deceased are fixed, and their interpretation is, in a sense, imposed by the individual (Davis, 2007). In contrast, the information in dialogue with deathbots emerges from an anticipated possibility—it is uncertain and hypothetical. Through this dynamic, ongoing updating, the construction of the deceased aligns with Levinas’ assertion that “through the experience of death we encounter the otherness of the other person.”

The dilemma of deathbots in the philosophical practice of death: analgesia, dignity of the deceased, and the future

The double-edged sword effect of deathbots in alleviating grief

Although deathbots have alleviated our grief in the farewell to death (Beischel et al. 2014), soothed our death anxiety, and shaped our dialogue with death in an entirely new way (Henrickson, 2023), yet the emotional entanglement with deathbots remains a cause for concern. Multiple studies have expressed concerns about the intimate relationship between humans and deathbots (Bao and Zeng, 2024; Fabry and Alfano, 2024). The greatest risk in interacting with deathbots does not lie in the inability to establish emotional connections with them, but rather in the self-harm and social dysfunction that can arise from an excessive obsession with these intimate relationships.

Just as with criticisms of internet addiction (Chou et al. 2005), an excessive focus on virtual relationships can lead to social problems in reality and other health issues. Fundamentally, deathbots are still only a technologically mediated, creative form of communicating with the deceased (Jiménez-Alonso and Brescó de Luna, 2023), an imagined, simulated dialogue with the deceased (Cholbi, 2021). The essence of this relationship is not based on the real resurrection of the deceased, but rather on a simulation of the deceased: a pre-set dialogue of possible interactions. Within this context, deathbots can only have two types of social relationships: one as a reflection of the deceased’s existence, and the other as a relationship with the interlocutor. The positioning of deathbots is very similar to the rumination thought processes of individuals after the breakdown of an intimate relationship (Saffrey and Ehrenberg, 2007). After a relationship ends, individuals will constantly engage in hypothetical dialogues in their minds, and these intrusive thoughts born out of the relationship’s dissolution can bring more painful experiences (del Palacio‐González et al. 2017), negatively impacting various dimensions of the individual’s mental health (Watkins and Roberts, 2020).

Although the positive therapeutic effects of deathbots in alleviating grief have been mentioned, in interactions with deathbots, maintaining an imagined intimacy can have potential negative consequences. From Heidegger’s perspective, as a product of technology, deathbots could easily obscure our authentic existence, potentially trapping us in das Mann-style inauthenticity. This inconsistency in outcomes highlights the two-sided of deathbots as an affective scaffolding. The reasons for this divergence may exist in two aspects. On one hand, besides the influence of individual traits on how one cope with the end of a relationship, whether deathbots become an addictive form of analgesia or a good buffering mechanism depends more on the micro-level interactions between the individual and the deathbots. In other words, the individual’s quest for dialogue may be either soothing in nature or a search for meaning. It is difficult to guarantee that the conversation with deathbots will necessarily reveal meaningful insights. On the other hand, the individual’s particular stage of grief, the deceased’s characteristics, the level of emotional disclosure during the interaction, and the focus of the dialogue process can all potentially steer the conversation in unpredictable directions. Therefore, in the interactions with deathbots, the involvement of mental health professionals and philosophical practitioners is necessary as a form of support, to help avoid the interactions from leading to outcomes that endanger the individual.

The existence of the deceased in deathbots: relationships, dignity, and autonomy

Relationships always have psychological and temporal boundaries. Even if the interactions in a relationship do not end due to death, they will gradually recede or enter a stabilized phase for various reasons. However, the boundaries of the relationship with deathbots are very difficult to fit into any relationship model, as it is a relationship where a material other and a spiritual subject overlap. In the relationship between the living and the deceased, after a period of exploration following the death of the deceased, the individual will ultimately incorporate the deceased into a symbolic form of existence, thereby stabilizing the relationship with the deceased through symbolic interactions (Nations et al. 2017).

Deathbots, however, are likely to prolong this exploratory period. Moreover, as a continuous dialogue, we cannot predict how this relationship should be terminated. As a simulation, deathbots will lose their meaning outside of the death event, whether in the short or long term. For the individual, the timing when the interaction with deathbots must end is inevitable, which means that the relationship with deathbots may face loss again. If the impact of the death event on the individual cannot be effectively processed within this relationship, secondary harm to the individual may occur (Bao and Zeng, 2024). When choosing how to confirm the stabilization of the relationship, we cannot ultimately transform it into a relationship between a person and an object, nor can we easily transition it into a human relationship. In the final process of dealing with this relationship, we must inevitably separate its material and spiritual attributes, thereby gradually integrating deathbots into the social course of the individual’s development. Although we can shape and control a relationship, as an interaction, it inherently comes with limitations. Excessive immersion in such interactions inevitably leads to what Heidegger refers to as inauthenticity (Lagerkvist, 2017).

Conversely, as a subject within social existence, a typical relational progression inevitably encompasses hesitation, withdrawal, silence, and various other obstacles. It also includes negative aspects such as anger, fear, and destruction, which are inherent elements of relationships. Thus, the depth and quality of interactions are temporally characterized (Sprecher and Hendrick, 2004). However, the deathbots appear as dead persons imprisoned by time, in which they are devoid of temporality and the emotional subjectivity of a living being, the emotion expressed in the dialogue is essentially the calculation of the dead, and is a kind of “second-hand” information. Although deathbots are constructed from the deceased’s past personality traits, data, and personal information, it is the individual (the living) who drives the change in the relationship during the conversation. Consequently, within this interaction, deathbots become mere adjuncts to the relationship. Because of the absence of agency, the dialogue is confined to a specific space in which there is neither temporality nor sociality. “When we text our friends and family, we may reasonably expect delays, forgotten responses, and updates with unexpectedly sudden personal news. The thanabot cannot provide this; a thanabot’s corresponding deceased person has essentially been reduced to a servile role with virtually no individuality or agency” (Henrickson, 2023). The deceased are “frozen” in a realm distinct from both the living and the dead. In this context, deathbots do not “resurrect” the deceased but render them “deathless”. Therefore, from the standpoint of the deceased’s dignity, the existence of deathbots itself still challenges the dignity of the deceased as a human being, even without considering the issues of data privacy and informed consent (Lindemann, 2022), and it shows a reduction of the personal right of the deceased during the interaction.

A post-human landscape of the deathbots

Technology, inherently devoid of directionality, is frequently moulded by social and cultural influences. We direct technology toward fulfilling the intrinsic desires of the human. Analogous to the role played by science and the scientific community in the emergence of nuclear weapons, the utilization of technology fundamentally reveals the aspirations within human nature (Hecht, 2010). Whether it involves the fear of destruction or the pursuit of power, humanity shapes technology and is not just shaped by it. The deployment of deathbots is an inevitable consequence of technological progress, driven by the profound human fear of death and the quest for immortality. According to terror management theory (TMT), human’s desire for immortality and the reality of physical death constitutes the greatest tension in human nature, and the pursuit of symbolic immortality becomes a defense mechanism against death (Pyszczynski et al. 2015).

Deathbots serve as a symbolic existence of the deceased in this new era, and in this way appease our inner fear of death, so the pursuit of more precise, more personalized, emotionally nuanced and even more proactive deathbots becomes an irresistible temptation. Despite critiques that digital resurrection technology might reduce the deceased to “electronic pets” (Lindemann, 2022), it is undeniable that GPT-based digital resurrection fundamentally challenges our perceptions of death. This technology emerges not merely from a need for consolation from loved ones but from a natural progression facilitated by technology itself. Given the current self-learning capabilities and technological advancements of GPT, the likelihood of a “technological singularity” occurring in the foreseeable future is considerable (Wei et al. 2022). Additionally, with the use of DAN (Do Anything Now) jailbreak attempts (a method for bypassing official GPT output restrictions through specific prompts), GPT demonstrates characteristics driven by death (Shen et al. 2023). In the future, data simulations are likely to achieve a level indistinguishable from reality. In such a post-human era, resurrecting the deceased might transcend mere simulation and become a form of real resurrection. In this envisioned future, how should we understand death and existence?

The digital existence of the deceased fundamentally overturns previous ontological thinking, representing a paradoxical phenomenon in existence (Lagerkvist, 2018). However, the phenomenological perspective of existentialism can offer a foundational basis. In Merleau-Ponty’s theory, temporality serves as a “field of presence” (Mensch, 2021), meaning that temporality endows spatiality with a characteristic of existence (Mensch, 2020). This definition of existence and identity based on Spatio-temporal continuity is used to explain the Ship of Theseus paradox (Yavuz, 2017). Similarly, we can use it to explain deathbots in the post-human era. Even if digital technology achieves a physical revival of the deceased to a completely consistent degree, the existence of the deceased is based on specific times and spaces. And from a broader social existence perspective, the identity of the deceased is not only defined by internal characteristics, but its experience embedded in specific Spatio-temporal relation constitutes the uniqueness of the existence of the deceased. Therefore, even the deceased (deathbots) who are exactly the same, are still another existence. However, this explanation is based merely on our phenomenological exploration, and when that vividly resurrected loved one stands before us, the complex emotions and astonishment involved will inevitably trigger psychological and philosophical confusion.

Epilogue

Technology, with its indifferent and malleable nature, is a mirror that reflects our deepest desires and fears, including our quest for control over life and death. Deathbots, as an extension of this technological progression, underscore the paradox of our mortality and the longing for symbolic immortality. As Borges describes in The Immortal, the cave dwellers, upon stepping into the river of eternal life, achieve immortality. They experience all possible events, both theoretically and potentially. For them, the concept of possibility becomes a fallacy, because all possibilities become inevitable (Mohammad Rezaee and Mehrianpoor, 2020). As a result, experience and existence themselves lose their meaning. This story can be seen as a metaphor for the relationship between technology and death and serves as a stark reminder that our relationship with death is not static but evolving, and in this evolution, we are not merely passive recipients but active participants shaping the conversation about what it means to be alive and what lies beyond.