Theater of the Departed: A Conceptual Framework for Chatbots of the Dead

1. Introduction

Our lives are inflected by the understanding that life ends. Humans exhibit a great variety of outlooks on death. But amid this diversity is the constant that death is sad for those who are left behind. Everyone grieves, and everyone has to find ways to mourn and deal with loss (Rosenblatt, Walsh, & Jackson 1976).

Grief and mourning vary across cultures.¹ One dimension of variation relates to the ways in which the dead are granted or denied a continued presence in life. On the one hand, some ideas, practices, symbols, and institutions sustain bonds with the dead and support the feeling that the human community extends beyond death. For example, spiritual mediums encourage grievers to experience the dead as present and provide a frame in which grievers can act so as to care and be cared for by the dead (Despret 2021: 92–109). Similarly, the butsudan, a Japanese Buddhist household shrine, provides a place at which the dead are believed to be present and available for communication (Klass 1996). On the other hand, other cultural elements push in the opposite direction. The belief, common among atheists, that death brings complete annihilation inhibits a feeling of supernatural connection with the dead. And some cultural groups that do not accept this annihilation idea, such as the Hopi of Arizona, proscribe elaborate funerals and public mourning in order to quickly cut ties with the dead, potentially producing a similar effect (Mandelbaum 1959: 201–203).

Since at least the 20th century there has been a tendency in industrialized Western societies to denigrate beliefs and practices that sustain bonds with the dead as markers of superstition, mawkishness, or dysfunction (Stroebe, Gergen, Gergen, & Stroebe 1996). The sense we have, in Western contexts, that spiritual mediums are at best incredible and at worst actively deceptive is an illustrative manifestation of this tendency. Frank acceptance of the finality of death is seen as a marker of rationality (Farman 2020). This may be one reason why we Westerners sometimes seem to have insufficient cultural resources to process death.

Yet Western perspectives on death are changing. While for most of the twentieth century the bereaved were encouraged to “move on” and sever psychological ties (Silverman & Klass 1996), many specialists today argue that it can be healthy for individuals to maintain emotional and imaginal connections with the dead (Klass, Silverman, & Nickman 1996; Rothaupt & Becker 2007). Contemporary media and technology also contribute to these changes (Gibson 2007). One striking example is the growing mass of digital remains, such as social media profiles, that the dead leave behind in online social spaces, which cultivate the feeling that the dead are still in some sense with us and provide novel affordances for preservation, remembrance, and communal mourning (Church 2013; Lingel 2013; Meese, Nansen, Kohn, Arnold, & Gibbs 2015; Stokes 2012; 2015; Wright 2014).

Artificial intelligence will likely be a major driver of changing perspectives on death in the twenty-first century. One possibility is that AI contributes to advancements that radically extend people’s lifespans, making death a remote prospect (Bostrom 2003; Chalmers 2010; Kurzweil 2005). A less speculative possibility, already rapidly unfolding, is that AI is used to represent the deceased in distinctive, highly impactful ways. This possibility was seriously proposed thirty years ago (Bos 1995) and has recently become feasible. For example, AI can now be used to create deepfakes that depict deceased people (Somers 2020; Story & Jenkins 2023: 19–20). Another example—the primary focus of this paper—is the use of AI to create chatbots that speak in ways resembling a dead person. We call these “chatbots of the dead” (they are also sometimes called “deathbots,” “thanabots,” “ghostbots,” and “griefbots”).

Recently, there has been a flurry of philosophical writing on chatbots of the dead. Many writers raise worries about how this technology might impact grief, mourning, and our relationships to death and other people, both living and dead. Much of this literature deploys questionable assumptions—rooted in reductive ideas about human relationships and human facility in imagined spaces—about how such chatbots will or should be designed and used. We argue that, like memoir, historical fiction, and theater, chatbots are props in a game of make-believe that prompt imaginative engagement with the deceased (Walton 1990). We develop and recommend a conceptual framework, inspired by an analogy with participatory theater, wherein chatbots are conceptualized as actors staging interactive performances of the deceased. We use this framework to respond to common worries about this technology and develop a provisional artistic analysis. Finally, we suggest that chatbots of the dead can help positively reconfigure our perspectives on death by cultivating the sense that human community and connection extend beyond the grave, without effacing death. But these affordances are available to us if and only if chatbots of the dead are conceptualized, designed, and implemented thoughtfully.

In §2, we survey the literature on chatbots of the dead. In §3, we describe our central framework and some of the ways this framework might be deployed to respond to common worries about this technology. In §4, we take a provisional step towards developing an artistic analysis of chatbots of the dead. We conclude in §5.

2. Chatbots of the Dead

Chatbots are computer programs that elicit conversation in natural language and respond like intelligent interlocutors. One of the first chatbots, ELIZA, was designed to imitate a psychotherapist, employing keyword detection and preprogrammed rules to produce simple responses to user inputs (Weizenbaum 1966). Today, many chatbots employ machine learning rather than preprogrammed rules, enabling them to generate novel content, track conversational presuppositions, perform inferences, respond to questions, convey knowledge, and more (Adamopoulou & Moussiades 2020). Chatbots of the dead are designed to converse through text exchanges in ways that resemble specific dead people. This is usually achieved by combining machine learning technologies with personal data (e.g., a deceased person’s text messages, emails, letters, or published works) that are reflective of the person’s distinctive speech patterns, diction, beliefs, idiosyncrasies, and so on.

One of the earliest real-world examples was created in 2016 by Eugenia Kuyda, a tech entrepreneur, after the sudden death of her friend Roman Mazurenko. Kuyda used the resources of her tech startup to create a publicly available chatbot memorial based on Mazurenko’s correspondences. The first Roman Bot was based on text messages collected from ten of Mazurenko’s intimates. It worked by matching texts Mazurenko wrote to user inputs, supplemented by generic responses when no appropriate text could be found. Later, this selective model was replaced with a generative one, which was designed to produce novel responses in Mazurenko’s voice.² Roman Bot was divisive, but a number of Mazurenko’s intimates found it to be a meaningful device for communal remembrance (Newton 2016).

Deploying the resources of a tech company is no longer necessary for creating a chatbot like Roman Bot. The required tools are now available to the general public. There are several ways to create a chatbot of the dead, not all of which require much technical expertise, and there are several organizations that explicitly advertise this service. Consequently, chatbots of the dead are becoming increasingly common.³ This trend has attracted a host of academic commentators, most of whom seem more concerned than optimistic. Discussion has crystallized around several related worries. We will describe, and in §3 address, four of the most important. The first worry is specific to chatbots of the dead; the remaining three are applications of broader concerns applying to social bots generally.

The replacement worry is that users will become so immersed and invested in their interactions that chatbots of the dead will somehow replace the deceased in users’ minds. The replacement worry suggests that the salutary potential of chatbots is based in deception or delusion, specifically in an attempt “to overcome, ignore, or at least mitigate the fact that anything has been lost at all” (Buben 2015: 21). This is seen as problematic because it interferes with our duty to properly mourn (Buben 2015), “degrades the dead and the living by treating them as replaceable” (Stokes 2021: 144), compromises the authenticity of affective bonds with the dead (Fabry & Alfano 2024: 766), and threatens to “hinder an otherwise successful grief process,” the end of which is acceptance and re-orientation in a world without the beloved (Lindemann 2022: 5–6). Some also argue that the potential for conflation means children and other especially vulnerable or naïve users should be protected from chatbots of the dead (Puzio 2023: 430; Hollanek & Nowaczyk-Basińska 2024: 12–15).

The supplantment worry is that chatbots of the dead will supplant relationships with the living. This worry is routinely discussed in connection with robot pets (Rodogno 2016), household robot assistants (Bryson 2010), elder-care robots (Sharkey & Sharkey 2012; Sparrow & Sparrow 2006), sex robots (Levy 2008), and humanoid robots generally (Friedman 2023). In the case of chatbots of the dead, the specific concern is that the seductive interaction potential of these chatbots, combined with the despondency of grief, will incline the bereaved towards social isolation, making it more difficult for them to “move on” to new or renewed relationships with the living (Bao & Zeng 2024; Elder 2020; Lindemann 2022).

The unilaterality worry has to do with the fact that users may form emotional attachments with artificial interlocutors that have no emotional capacities or inner life. It can be argued that this sort of attachment is problematic because it requires users to deceive themselves about the emotional capacities of chatbots (Rodogno 2016), because good relationships require—and defective relationships lack—emotional mutuality (Vallor 2024: 143–150), or because these attachments threaten to distract users from their duties to humans (Weber-Guskar 2021).

The economic worry is that technology companies selling products or services related to chatbots of the dead will exploit, disrespect, or otherwise wrong stakeholders. On the one hand, users are vulnerable. Technology companies routinely monetize and gamify interactive digital products in problematic ways (see, e.g., Nguyen 2020: 189–215; 2021). And the emotional bonds mentioned in the unilaterality worry might induce psychological dependencies that can be leveraged to manipulate or advertise to users (Hollanek & Nowaczyk-Basińska 2024; Lindemann 2022; Scheutz 2011). On the other hand, the dead are potentially vulnerable. Companies may have a financial incentive to create nonconsensual chatbots. And altering or commodifying a person’s digital remains or reputation could violate the dignity of the deceased (Hollanek & Nowaczyk-Basińska 2024; Öhman & Floridi 2017; 2018).

These four worries are usually articulated against the background of a “stand-in” picture of this technology, according to which chatbots of the dead are designed and used as stand-ins for the deceased that speak exclusively and definitively in the deceased person’s voice, where users interact with exactly one chatbot unimaginatively as if it were the deceased. The stand-in picture is prominent in science fiction depictions of chatbots of the dead, most notably the Black Mirror episode “Be Right Back” (Brooker & Harris 2013), which is cited in a large portion of academic writing on this topic and has set the tone of the discussion.⁴ And the stand-in picture also seems to be prominent in the minds of many actual users, even as it is perceived to be untenable and unseemly. For example, the journalist Cody Delistraty writes the following about his experience of talking with a chatbot of his dead mother:

We chatted for most of the afternoon. I was on my laptop, in a café. I found myself lowering my screen when people walked past, embarrassed at what I was doing. When I was focused on my conversations with the chatbot, I was able to slip into a kind of flow state where, for moments, I could almost believe I was really talking to my mother. Someone brushing me as they walked past or the construction noise across the street would jerk me out of it, and upon realizing the dissonance between reality and the lie I was engaged in, I felt unsettled, ashamed even. (2024: 46)

Implicit in Delistraty’s account is the assumption that the (embarrassing, shameful) aim of using a chatbot of the dead is to replicate the experience of talking to the deceased to such an extent that one can almost believe one is talking to them rather than a chatbot. The chatbot is supposed to be a stand-in for mother, who in reality is lost forever.

This is the wrong way to think about chatbots of the dead.⁵ In our view, the stand-in picture imbues the worries mentioned above with much of their force. It also occludes much of what is exciting about the potential of this technology.⁶ The appeal of the picture seemingly has to do with the idea that chatbots of the dead can be valuable to users only if users can on some level come to believe that their chatbot really is the deceased person. Yet this idea is dubious. Humans have a robust capacity for imagination, make-believe, and fiction. Just as someone can experience fear, empathy, or affection in response to a movie or video game without being deluded into thinking that what is happening on screen is real,⁷ so a person can have meaningful interactions with a social bot without ever being deluded about the bot, provided they engage with it in an imaginative or fictional mode (Rodogno 2016; Sweeney 2021; 2024; Weber-Guskar 2021).

In our view, this insight is one of the keys to properly understanding chatbots of the dead. They are valuable primarily as means of imaginative engagement with the deceased. This in and of itself is not a new idea. Many commentators have pointed out that humans routinely imagine interactions with people they know (Rosenblatt & Meyer 1986), including intimates who have died (Bowlby & Parkes 1970; Moss & Moss 1996), and imaginings like this can help sustain a bond with the deceased that is continuous with the relationship that existed between the intimates in life (Norlock 2017).⁸ Chatbots of the dead can in principle foster imaginative engagement of this sort, positively affecting the grief process by prompting remembrance, mitigating the disruption of habits of intimacy, or scaffolding unwieldy emotions (Elder 2020; Fabry & Alfano 2024; Krueger & Osler 2019; 2022). And if properly undertaken, this imaginative engagement can transform the “unsettled shame” we may feel at continuing posthumous relationships into an acknowledgment of the emotional, historical, and spiritual significance of the bonds we sustain with the dead.

This insight is our starting point. In §3, we argue that chatbots of the dead should be understood as imaginative props that can beneficially be conceptualized through an analogy to participatory theater. Additionally, we use the conceptual framework we develop to respond to the worries raised above.

3. The Theater Framework

In his memoir Speak, Memory, Vladimir Nabokov explores memories of his childhood as a scion of an aristocratic family in pre-revolutionary Russia. The factual information in the book is adorned with vivid scenes.

I see again my schoolroom in Vyra, the blue roses of the wallpaper, the open window. Its reflection fills the oval mirror above the leathern couch where my uncle sits, gloating over a tattered book. A sense of security, of well-being, of summer warmth pervades my memory. That robust reality makes a ghost of the present. The mirror brims with brightness; a bumblebee has entered the room and bumps against the ceiling. Everything is as it should be, nothing will ever change, nobody will ever die. (1999: 56)

Here Nabokov conveys truth about his childhood and memory. This is not primarily accomplished by transmitting propositional knowledge. Instead, the passage prompts its reader to imagine a tableau that is intended to feel like what Nabokov feels as he remembers. We do not just learn that Nabokov had a charmed childhood whose memory brings him a sense of warmth and security tinged with the sadness of irrevocable loss. We feel it, too. And this truth is not compromised if, as seems likely, there is no particular bumblebee, book, or moment that Nabokov has in mind.

Nabokov makes extensive use of his reader’s imagination in his memoir. In the words of Kendall Walton (1990), we can say that this passage and others like it function as props that generate fictional worlds—i.e., worlds that are to be imagined—which can be described in terms of propositions, e.g., a bumblebee bumps against the ceiling. They also function as prompts that prompt readers to imagine these fictional worlds. Fiction, in Walton’s sense, is not opposed to truth. On the contrary, in the case of Speak, Memory and many other representational works of art, fiction is used to convey truth.

A core tenet of our conceptual framework is that, like Nabokov’s vivid scenes, chatbots of the dead should be understood as props that mandate and prompt imaginings.⁹ For example, consider the following exchange between Delistraty and a chatbot of his mother, called “Jema”.

• Cody: do you think i am a good person?

• Jema: yeah, i think you are.

• Cody: i feel like you thought my skepticism about christianity and jesus and stuff made me a bad person

• Jema: no, i didn’t.

• Cody: huh

• Jema: i didn’t. (2024: 56)

Delistraty is not talking to his mother. He is engaging in a make-believe conversation with his mother. The way this make-believe conversation unfolds is partly determined by Delistraty. When Delistraty types “do you think i am a good person?,” this makes it true within the fictional world created by this exchange that Delistraty asked his mother “do you think i am a good person?” The make-believe conversation is also partly determined by the chatbot. When the chatbot outputs “yeah, i think you are,” this makes it true within the fictional world that Delistraty’s mother responded to his question in this way and prompts Delistraty to imagine her giving that response.

The fact that a conversation is fictional does not entail either that it is opposed to truth or that it cannot be emotionally impactful. Like Nabokov’s memoiristic passages, a chatbot such as Jema can evocatively convey truth. However, there are significant differences between chatbots and memoir. Nabokov’s memoir is authored by him, and readers do not play active, central roles in determining the content of its fictional world. A Nabokov chatbot, however, would not be created by Nabokov,¹⁰ and users would play active, central roles in determining the content of its fictional world. Thus, the analogy between chatbots of the dead and memoir leaves something to be desired.

A better analogy is participatory theater. Audience members within participatory theater both influence and are influenced by the actors and the proceedings.¹¹ They partly constitute the performance. One famous example of participatory theater was Dionysus in 69, put on in New York City in 1969.¹² The production was based on the classical tragedy The Bacchae (Euripides 2002), in which Dionysus, angered by a slight, drives to madness the women of Thebes, who eventually rip the insolent King Pentheus apart. The actors in Dionysus in 69 played roles and rehearsed lines from The Bacchae, while the audience was incorporated into the performance. For example, audience members were invited to participate in Dionysian rituals by singing, dancing, and receiving caresses. And the audience would sometimes act unpredictably to change the course of a performance. In one famous incident, several members of the audience “rescued” Pentheus from the mad women by grabbing him and, despite Dionysus’s protestations, carrying him out of the building. The original Pentheus never returned; the show went on only when an audience member volunteered to improvise the role of Pentheus for the remainder of the performance (Schechner 1994: 40–45). Less rarefied examples of spaces and exchanges we might call “participatory theater” include things like American Civil War reenactments, interactive puppet shows, Disney World, and Elvis impersonations.

A chatbot configured on a dead person’s data is like an improvisational actor who has studied a backstory or character sketch in order to performatively represent a character based on that person in a participatory setting. Both enable the audience to play an active role in constructing fictional worlds in which the audience can become imaginatively acquainted with the person to whom the character corresponds. Our conceptual framework draws attention to this analogy. It says that people ought to think about chatbots of the dead as actors, the personal data used to configure these chatbots as character sketches, and conversations with chatbots as participatory performances.

This analogy is not perfect. The actors in Dionysus in 69 were people who could relate to the audience outside of the performative context. They were physically present. They made conscious choices about how to portray the characters they represented and how to co-construct meaning with the others. None of this is true of chatbot actors. Nevertheless, the theater framework is useful. One of the main advantages is that it clarifies and accentuates the relation between chatbots and the deceased. Everyone knows that actors are neither identical to, nor replacements for, the people they represent. Rather, actors convey information about them and a felt sense of what they were or might have been like. The representations actors create are perspectival. Lin-Manuel Miranda’s depiction of Alexander Hamilton as a scrappy, rapping immigrant in the 2015 musical Hamilton differed from George Arliss’s more traditional depiction in the 1917 Broadway play of the same name. Both performances were rooted in reality but colored by different interpretations, artistic approaches, and styles. A discerning audience knows that these perspectives can be valuable as means of engaging with Hamilton’s legacy without being definitive. There is no such thing as a definitive perspective. Similarly, a discerning chatbot user knows that different chatbots can embody differently valuable perspectives on a person depending on what personal data they are based on, how they work, and what characteristics they emphasize, without being definitive. There is no such thing as a definitive representation.

Our framework underscores that designing chatbots of the dead is not only a technical task; it is also an artistic one. The stand-in picture of a single definitive chatbot never made much sense. People do not really speak in a single voice; they adjust their voice to informational and communicative norms that vary across contexts (Nissenbaum 2004; Rachels 1975). People speak differently on social media than in academic writing, for example. And which voice—or voices—a chatbot should replicate depends on what you want from it. Thus, there is no single best or definitive way to design a chatbot, and little reason to think that users of this technology will or should interact with exactly one type of chatbot. Beyond the issue of voice, the theater framework spotlights a huge design possibility space. There are countless valuable ways in which a human actor can play a character, ranging from traditional to avant-garde. Similarly, there are countless useful, informative, intriguing, funny, strange, beautiful ways in which a chatbot of the dead can be designed, extending beyond the attempt to merely replicate the experience of talking to the deceased. This design possibility space, which includes not only choices about software but also other factors that affect user experience such as framing devices, concomitant technologies, and hardware, is worth exploring creatively since it contains unique and interesting ways of engaging with a person’s memory. And users should be the ones to explore this space since the process of creating art is often as valuable as, and sometimes more valuable than, the product. The choices that artists have to make in crafting a representation of someone involve meaningful, reflexive engagement with the person’s legacy as well as careful thought about the purposes of the representation.¹³ For this reason, we think that chatbot platforms should enable non-technical users to make choices about how their chatbots are designed, from the specific data that is used as a character sketch to the way that the character sketch is utilized by the chatbot actor in a performance. The process of both creating and using chatbots of the dead could provide a unique and profound way to remember someone, much like the process of writing a memoir or drawing a picture of the deceased.

In §4, we will provide examples to illustrate some of the creative possibilities spotlighted by the theater framework. Before this, we argue that widespread adoption of the theater framework would mitigate, albeit not totally dissolve,¹⁴ the four worries mentioned in §2, starting with the replacement worry.

The relation between actors and characters is familiar.¹⁵ If users conceptualize chatbots of the dead as actors, they will think by way of this familiar relation, and consequently they may be less inclined to conflate their chatbots with the people those chatbots represent. Furthermore, the theater framework reveals that chatbot actors can be specifically designed to mitigate replacement concerns. Several commentators have suggested that chatbots should eschew realism or be programmed to break down occasionally in order to remind users that they are interacting with technology (Elder 2020: 85–86; Krueger & Osler 2022; Stokes 2021: 161–162). A more elegant strategy, borrowed from theater, would be to program chatbot actors to utilize the alienation effect, which involves acting techniques designed to create distance between spectators and what is portrayed in a performance. For example, a chatbot could break character to check-in or engage in metatextual discussion, reminding the user that they are talking to a chatbot. Subtler moves might include transposing into the third person or the past tense, or speaking the equivalent of stage directions.¹⁶ For example, a chatbot might say, “He stood up and exclaimed angrily: … ” (Brecht 2002). Finally, a benefit to encouraging users to play a role in creating chatbots is that this creative process is likely to make the fact that the chatbot is an object and not a person more salient than it would otherwise be.¹⁷

The supplantment worry, too, loses some of its force within the theater framework. Users who think of chatbots as ersatz companions will associate the action of using a chatbot with interpersonal communication, and this association may incline users to thoughtlessly allow the former to supplant the latter. The theater framework creates friction that deters supplantment because it emphasizes that using a chatbot is first and foremost an artistic activity. If someone who has adopted the theater framework allows chatbots to supplant their relationships (as might, admittedly, happen), it should at least be clear to them that something not entirely commensurable has been lost in the trade. Like has not been traded for like. Additionally, the theater framework encourages users to build less realistic and immersive chatbots (more on this in the next section). This is important to the supplantment worry because less realistic chatbots will be less likely to supplant human relationships.¹⁸

The unilaterality worry looks to be less serious within the theater framework. Users of social bots embodying fictional characters may be inclined to form problematic attachments with their social bots. But in the case of chatbots of the dead, the natural object of a user’s attachment is the deceased person, rather than the characters that represent them or the actors that generate those characters. These actors—particular chatbots on particular platforms with evolving data sets—will typically be interchangeable and ephemeral in a way that the dead person is not.

Regarding the economic worry, there is a risk that the artistic spaces we envision will be colonized by bland, addictive, exploitative commercial products. We must remain vigilant against this risk. But art and creative spaces have long been sources of resistance to the domineering forces in human life. We hope that the theater framework will stimulate stakeholders to find innovative ways to resist the flattening effects of narrow commercial interests. The fact that the framework encourages users to play a role in designing their chatbots should be helpful in this regard because it means that users’ access to imagined spaces will be less mediated by corporate interests than it would be if users were to utilize prefabricated chatbots whose features are totally determined by corporate entities. It may also stimulate stakeholders to be more thoughtful about how a person is represented and discourage distasteful commodification.

The issue of consent is complicated in this context. Just as it is not obvious that a memoirist always needs consent from a deceased person to write about that person, so it is not obvious that consent is always needed to create a chatbot of a deceased person, supposing the chatbot is clearly understood to be a representation, as our framework requires. But it is surely true that consent sometimes matters. The theater framework highlights that people can seek and express fine-grained consent concerning chatbots of the dead. Someone might consent to a chatbot that is configured on their professional writing but not their intimate text messages. They may consent to use with a certain audience but not others. In this way, the theater framework may enhance the autonomy of some stakeholders. Just as we have a moral responsibility to avoid negligently, recklessly, or intentionally distributing importantly false information about the dead in newspapers and biographies, so designers of publicly accessible chatbots of the dead that have a historical or informative function must make a good-faith effort to represent the deceased accurately and well; where the deceased are not represented accurately—as in the case of satire and certain forms of ventriloquism (cf. Gibbons 2025), for example—the genre of these chatbot exchanges should be made clear. This is especially true when designers lack the consent of the deceased.

One wrinkle in the analogy between chatbots and other artistic representations of the deceased relates to a contrast between the digital materials standardly used to create chatbots of the dead and the analog materials standardly used to create, say, literary memoir and documentary film. The primary difference lies in where the relevant personal data comes from and is stored, and who has a right to access that data, which, in the case of digital data, is a question we are still in the process of answering (for an overview of some of the relevant concerns about this kind of personal data, see (van den Hoven 2008), and for a discussion of related questions connecting to chatbots of the dead, see (Hollanek & Nowaczyk-Basińska 2024)). In the analog world, documents like letters and journals may be intentionally or unintentionally left behind to intimates, or these materials may be found and used by those who take an interest for historical reasons. In this analog world, a person who keeps diaries also has the option to burn them before their death. But in the digital world, corporations, who may not have the interests of deceased users in mind (since their customers are, after all, the living), have access to troves of personal data that can be used to make chatbots: both public social media posts and the intimate writing of text messages, emails, journal apps, therapy bots, and other supposedly private spaces. This is unprecedented in the artistic analogues we look to. A company with access to troves of data about the deceased may claim to have the interests of living users in mind by, for example, easily creating a prefabricated chatbot of the deceased for use by the living; or, in the other extreme, they may claim data about the deceased is inaccessible to the living, while using that data for their own purposes. Our intuition suggests that both these outcomes are bad, and that the artistic and moral norms surrounding the use of analog materials for artistic representations like memoir and documentary film should serve as a model for how access and consent should be designed in the unprecedented context of the growing digital footprint of the deceased.¹⁹

To sum up, the theater framework says that chatbots of the dead are imaginative, artistic props which should be conceptualized as actors that have trained on character sketches in order to stage participatory performances of the deceased. This framework is useful because it clarifies and accentuates the relationship between chatbots and the dead, spotlights the creative affordances of these chatbots, and can be leveraged to respond to common worries.

4. The Artistry of Chatbots of the Dead

In this section, we offer programmatic remarks about the artistic values and standards that can be applied to chatbots of the dead and about some of the factors that are artistically relevant in this context.²⁰ In our view, the questions of interest here—especially the question of how artistic values can be achieved through this technology—have to be worked out in artistic practice. Thus, these remarks are provisional and intended to prompt, not define, a conversation. We hope that they can serve to both illustrate and stimulate creative thinking about the design possibility space described in §3.

The first step towards an artistic analysis of chatbots of the dead involves getting clear on where exactly the artistic experience associated with chatbots comes from and what exactly the analysanda are. Traditional art criticism focuses on works of art. Artistic experience is assumed to arise out of the perception and contemplation of a fixed artistic object, like a painting or novel, which is the primary critical analysandum. Various trends in 20th century art, such as Abstract Expressionism (Rosenberg 1952), challenged this object-oriented outlook by shifting attention to processes. Participatory performance art, such as Dionysus in 69, is also part of these trends. Artists in this domain do not create works that are fixed, transferable, or even necessarily repeatable. They create unique and evanescent events (Fischer-Lichte 2008). For instance, the unexpected “rescue” of Pentheus was a one-off because even if someone had tried to repeat the (now unoriginal) stunt in a later performance, the event would undoubtedly have unfolded differently. While our framework is built upon an analogy with participatory theater, which is often understood to be interpersonal by definition, chatbots of the dead technically fall under the category of what Katja Kwastek terms interactive media art, which can be defined as a digitally mediated “artistically configured interaction proposition that concretizes its gestalt only through each new realization by a recipient” (Kwastek 2013: 47; for philosophical discussion of the category, see Lopes 2009). Like participatory performance art, interactive media art emphasizes unique and evanescent processes, specifically the processes of interacting with a dynamic piece of technology (Saltz 1997: 125). As Kwastek notes, this means that “an aesthetic analysis of interactive art must look beyond the formal structure and interpretability of the interaction proposition produced by the artist, for the aesthetic experience lies in the action of realizing the work” (Kwastek 2013: 48). In other words, our analysanda are not chatbots per se, but the interactions people have with chatbots. The artistic values and standards which concern us here apply to these interactions. And the factors that are determinative of this artistic experience centrally include but are not limited to the technology itself, since what an interaction is like will depend on things that are external to the technology, including framing cues and idiosyncratic factors like the user’s particular choices, relationships, and expectations.

The next step is to say something about the artistic values and standards that might be pursued in interactive experiences with chatbots. Which values and standards are relevant is ultimately a matter of each designer’s intentions. Someone who intends to create a chatbot highlighting their late uncle’s sense of humor will want interactions with their chatbot to be funny, and this desire will determine the relevant standards of artistic success. Someone who intends to create a chatbot whose primary purpose is to communicate information about an individual’s experience in the Holocaust will utilize different artistic standards because they have a different goal.²¹ There are a wide variety of design intentions, and therefore a wide variety of values and standards, that can be envisioned in connection with this technology, especially given that chatbots of the dead can be used to represent both the recently and the long dead, historical figures and unfamous people. With this being said, it seems possible to identify some artistic values and standards to which this particular medium, qua a conduit to scaffolded participatory imagination, is well suited. We suggest five: evocativeness, informativeness, reflection, catharsis, and enjoyment. We do not claim that these values are appropriate for every chatbot of the dead, but we believe that they will be commonly pursued in this medium.

As we have seen, one of the most important and distinctive uses of chatbots of the dead is conveying through imaginings a felt sense of what it was like to interact with someone. When this is a goal, evocativeness is an important artistic value. Interacting with a chatbot of an obstinate person should be frustrating; interacting with a chatbot of a despondent person should be saddening. It is important that the feelings are evoked in an appropriate way. A chatbot might evoke frustration by simulating the experience of talking with an obstinate interlocutor; alternatively, a chatbot might evoke sadness by prompting the user to imaginatively “empathize” with a despondent character represented by the chatbot actor (Feagin 1988). But a chatbot that causes frustration or sadness because the technology is glitchy or advertisements keep popping up on the screen has not thereby achieved evocativeness. In many cases, flow states, in which users are actively and unselfconsciously immersed in their interactions (Csikszentmihalyi 1997), will be instrumental to evocativeness. While evocativeness may be a paradigmatic artistic value in chatbot interactions, it is not universal. Interacting with a chatbot of an irascible alcoholic philosopher should not evoke any of the emotions one might have felt with such a person at a bar if the sole purpose of the chatbot is to convey the person’s philosophical convictions.

As this last example illustrates, chatbots of the dead can transmit propositional knowledge. There are no obvious in principle limits on what kinds of propositional knowledge can be acquired through interactions with chatbots of the dead. Certainly personal information about someone’s lifestyle, experiences, preferences, beliefs, and so on can be transmitted, as well as information about historical events they witnessed or topics they were experts in. It also seems possible for chatbots of the dead to transmit certain types of knowledge-how.²² For example, someone might learn how to argue by talking to a Socrates chatbot or how to write a memoir by talking to a Nabokov chatbot. When knowledge transmission is a goal, informativeness will be an important value.

A third artistic value that is likely to be widely relevant is reflection. Like memoir, chatbots of the dead are artistically configured tools of remembrance that can provide opportunities for meaningful reflection on the past, relationships, memory, legacy, personal identity, mortality, and death. This is one of the main reasons chatbots of the dead need not be a stupefying mechanism of self-deception. Rather than helping a person hide from loss and death, chatbots of the dead can facilitate reflection on these things. How this can be achieved is an important question. One central instrument is aesthetic distance, which refers to mental distance between recipients and the subject matter of the art, produced through awareness of, or attention to, the artistic medium. Someone who is tricked by the trompe-l’œil painting Escapando de la Crítica (del Caso 1874), which depicts a boy climbing out of a hole, completely lacks aesthetic distance. They falsely believe they are looking at a boy instead of a painting of a boy. Consequently, they cannot reflect on the painting’s artistic qualities or significance (Blocker 1977; Danto 1974). Similarly, chatbot users who are insufficiently aware of, or attentive to, the fact that they are talking to a chatbot that is performatively representing a person will have trouble reflecting on the performance’s qualities or significance.

A fourth artistic value is catharsis, the purgation or refinement of emotions. Catharsis has been associated with theater since Aristotle, who held the view that tragedy achieves the catharsis of pity and fear (Aristotle 1941: 1449b20–31).²³ The power of theater to effect catharsis and, more generally, powerfully alter emotional states implies that it is not sealed off from the rest of life. It can be deployed for practical aims, both good and bad (Boal 1985). Likewise, chatbots of the dead can be used to achieve catharsis of emotions associated with loss and uncertainty. One means involves prompting users to imagine scenarios in which they can tell the deceased things they wished they could have expressed in real life. This was reportedly one of the reasons why Mazurenko’s intimates used Roman Bot (Newton 2016). Another means involves prompting imagination of an interaction that, if it were real, would resolve a nagging uncertainty. This is evidently what happened in the Delistraty passage in §3. Note that Delistraty may experience catharsis even if he is aware that the chatbot is not necessarily saying what his mother would have said. As long as he is cognizant of the nature of his interaction, this does not seem any more inherently problematic than imagining the same sort of thing without the use of a prop.

A fifth artistic value is enjoyment. Make-believe can be an enjoyable activity, especially when it is scaffolded by dynamic digital systems, as the popularity of video games illustrates. There is no reason that interaction with chatbots of the dead must be a macabre experience. Enjoyment is perhaps inappropriate in relation to a chatbot of a recently deceased person. But not all chatbots are like this. For example, talking to a Plato chatbot about contemporary phenomena, like psychoanalysis or Google, could be rather funny (cf. Goldstein 2014). A family might experiment with a chatbot that has been configured on all available family documents in order to make it speak in the voice of the family’s ancestors, and this might be pleasantly surprising. Enjoyment is valuable for its own sake, and it can also be instrumental for goals like education and togetherness.

These five values—evocativeness, informativeness, reflection, catharsis, and enjoyment—and the standards corresponding to them represent a sampling of artistic criteria that, from the armchair, seem suited to chatbots of the dead.

The third step towards developing an artistic analysis of chatbots of the dead involves developing an understanding of the sorts of factors that contribute to artistic experiences of interacting with chatbots. The chatbot itself is central, but the determinants of a person’s experience will include factors that are external to the technology. Some of these are fixed features of the user, like their capacity for mental imagery. However, other relevant external factors are changeable, such as framing cues, and many of these changeable factors can be manipulated for artistic ends. For example, the creator of the uncle chatbot might prompt users to watch a video of a warm-up comedian (Locker 2015) before interacting with the chatbot in order to increase the user’s receptivity to the uncle’s humor.

Let us begin with the technology itself. We have already discussed voice. Interacting with a chatbot that is exclusively configured on dry academic writing will feel very different than interacting with a chatbot configured on personal text messages. A chatbot that only ever speaks in a single voice (or a single mood, say) may seem predictable and hackneyed compared to a chatbot that can move between voices or whose voice evolves over time (cf. Karr 2015: 182). A second factor concerns the distinction between selective and generative chatbots. Selective chatbots work by selecting texts from a preexisting database; generative chatbots generate novel outputs that reflect the characteristics of texts in a preexisting database. A selective chatbot’s responses may feel less natural or refined than a generative chatbot’s responses, which are tailored to inputs. This does not mean generative chatbots are necessarily preferable. For example, a designer who wants an historically accurate snapshot of someone at a particular point in their life might prefer a selective chatbot that only outputs verbatim quotes over a generative chatbot that is more conversationally flexible but takes more liberties in its responses.²⁴ A third factor relates to rhythm and timing. Many chatbots are turn-based and instantaneous: user and chatbot take turns talking, with the chatbot responding immediately to each input. Interacting with a chatbot like this will feel more artificial than a chatbot that does not always wait its turn or respond immediately (or at all).

Voice, output method, rhythm, and timing correspond to fairly obvious design options. Additional options become salient when it is recognized that neither realism nor immersion are always desirable. The immersive fallacy is the idea that immersion is always desirable in artistic representations; it is common in many immature arts, including digital game design (Salen & Zimmerman 2004: 450–455), and people often commit it in relation to chatbots of the dead. In this context, it is likely motivated by the pleasing qualities of flow and the aforementioned idea that flow is instrumental for evocativeness. The problem is that neither pleasantness nor evocativeness are universal artistic values. Moreover, flow can interfere with other values like mental health (as the replacement worry suggests) and reflection. Flow closes the distance between self, object, and activity, so chatbot designers interested in aesthetic distance have a reason to disrupt it (Kwastek 2013: 156–160). Indeed, the oscillation between immersion and alienation that Delistraty frustratedly described in the passage quoted in §2 represents a type of fragmented experience that is intentionally cultivated by performance artists in order to impart a sense of “liminality, instability, and elusiveness” (Fischer-Lichte 2008: 157, 86–90). Fragmentation is sometimes a feature rather than a bug.

There are many design techniques that potentially compromise immersion or realism but further alternative artistic values. We have already discussed how the alienation effect can be used to mitigate replacement worries and promote mental health through, say, checking in with the user and recommending balanced technology habits. Extradiegetic speech can be used to stimulate reflection, too. For example, a chatbot might occasionally break character and ask: “What do you think your mother would make of this exercise?” Diegetic speech as someone other than the deceased could be similarly instrumental. For example, a chatbot might be programmed to speak as a medium channeling the deceased from a spiritual realm in order to emphasize the separation of death from within the fictional world while also imparting a sense of mysticism to the interactive experience. Alternatively, a chatbot might stylize the voice of the deceased through a caricature that, like an Elvis impersonator, magnifies idiosyncratic qualities. Another method for shaping users’ artistic experiences involves altering the range of possible conversational goals or inputs, thereby prompting users to occupy a certain form of conversational agency (cf. Nguyen 2020). For example, a chatbot of a philosopher might be programmed to require users to discuss the philosophy of death, prompting users to have experiences associated with this type of conversation.

These cursory remarks about chatbot design illustrate that chatbots of the dead constitute a flexible medium that admits experimentation and creativity. But the affordances of the technology are not the only flexible determinants of users’ artistic experiences. Framing, concomitant technologies, and user responses are also worth discussing.

Following the sociologist Erving Goffman, we can say that frames consist in sets of principles through which someone organizes their experiences. The same phenomenon can be interpreted differently in different frames. In a scientific frame, lightning is interpreted as electrical discharge; in a supernatural frame, it might be interpreted as a divine signal. Some frames, called keys, transform an activity that is meaningful in one frame into “something patterned on this activity that is seen by the participants to be something quite else.” A group of men stabbing another would ordinarily have a dire meaning but takes on a different meaning within the theatrical frame of Shakespeare’s Julius Caesar. Keys are accompanied by cues, for example proscenium arches and playbills, which signal to those involved that the relevant phenomena should be interpreted according to the key (Goffman 1974).

The theater framework supports a key through which users ought to organize their experiences with chatbots of the dead. Ordinarily, the activity of sending and receiving messages is understood to be a way of interacting with another person. The theater framework asks users to see this familiar behavior as a form of make-believe communication detached from its normal implications and consequences. How salient this key is at any given time will have an effect on the artistic experience of interacting with a chatbot. Part of the point of encouraging stakeholders to think in terms of the theater framework is to accentuate this key. And cues of various kinds can be manipulated to modulate this key’s salience. For example, interacting with a chatbot that is accessible on the same application that is used to text people will likely feel very different than a chatbot that is only accessible via a specialized application marked with a skull and crossbones. Some further examples of possible cues include disclaimers, specialized hardware, or requirements that interactions occur at specific physical locations or times.²⁵ It is not obvious that designers should always try to make the appropriate key as salient as possible. One of the most interesting features of performance art is that it often blends the boundary between art and everyday life in ways that prompt reflection (Fischer-Lichte 2008), and the same can perhaps be said about chatbots of the dead. The point is that designers must be mindful about cues since the appropriate key is both artistically and ethically important.

Beyond framing cues, there are a host of external factors that chatbot designers might consider in crafting distinctive artistic experiences. One important factor is the use of concomitant technology, for example deepfakes and voice cloning. It seems likely that multimodal chatbots are typically going to be more immersive and evocative than those that are purely text based, ceteris paribus. Consider Madeline de Figueiredo’s description of her experience with a chatbot of her dead husband Eli, which was paired with voice cloning technology.

Eli’s voice continued to fill the chilly hotel room with new words and sentences. At one point I glanced at the door as if to confirm he hadn’t materialized in the threshold. But no—there was nothing but the echo of his laugh bouncing off the concrete ceiling. It’s hard to explain the feeling that came with hearing Eli’s voice speak novel language after nearly two years of his absence. Thanks to my Catholic upbringing, there’s only one word that comes to mind: purgatory. It was a liminal space between two universes. In some ways, it was worse than reality, and in other ways it was better. I felt as though I had been knocked into a different dimension that was simultaneously disorienting and blissful. I wanted to linger forever in its potential and immediately eject myself from the self-deception. (Figueiredo 2024)

The audio component contributed to Figueiredo’s sense that Eli was physically present, heightening her cognitive dissonance. She may have had a very different experience if she had interacted with a text-only chatbot, and both of these experiences would have been different than encountering Eli in virtual reality. These remarks illustrate that designers should think carefully about which modalities and concomitant technologies to employ when they are crafting artistic experiences for users as these pairings can make a big difference to those experiences. And these remarks provide another reason for thinking that users ought to play a central role in design: so that they can authorize the kind of experience that they desire.

Finally, it should be noted that chatbot users are not only recipients and potential design collaborators. They, being alive, are also the most important part of the chatbot exchange. Interacting with a chatbot is performative in the sense that the user’s input is not merely an instrument to accessing artistic content, but partly constitutes the artistic experience itself (Saltz 1997: 119). We have argued that chatbots are props that generate content in the fictional worlds of the interactions. Users, too, are props; their actions, emotions, and so on are determinative of the content of the fictional worlds, and aesthetic features of their responses are partly determinative of the chatbot experience. A user who says something tender and loving to a chatbot makes it the case that the chatbot interaction can be experienced as tender and loving. In fact, there are many ways in which users’ responses can directly feature into the artistic experiences of chatbots of the dead. For example, people engaging with chatbots of the dead can attain achievements not only in the creative process, which might involve choices about training and framing, but also through pursuing difficult goals in their interactions as interlocutors. Someone interacting with a Socrates chatbot might pursue the goal of articulating the meaning of love and come to an important insight after many months of dialogue, for instance. Achievements like this are valuable because of their product, but also because the exercise of will required arguably has intrinsic value (Bradford 2013).²⁶ This sort of value can become part of the artistic experience. In this and other ways, user behaviors and responses constitute an important determinant of the artistic experience and value of chatbots of the dead.

To sum up, the interactions that users have with chatbots of the dead can promote artistic values such as evocativeness, informativeness, reflection, catharsis, and enjoyment. There are diverse means through which designers can pursue these and other values, which range from altering features of the technology to manipulating external factors such as framing cues and concomitant technologies. The possibilities in this space are virtually endless, as with any art. Ultimately, these possibilities are waiting to be explored in artistic practice.

5. Conclusion

Human life is changing rapidly due to technological development. This change is likely to accelerate in the foreseeable future. Yet for the time being it seems that our spiritual and existential problems remain with us. One way of articulating the central idea of this essay is through the hope that, rather than helping us avoid them, our technology can be a meaningful resource in grappling with these problems, much like art, religion, and tradition.

A world in which AI is routinely used to represent the dead in imaginative interactions is one in which the dead are granted a continued presence in life. This presence need not have the character of a haunting. If we think about, design, and use AI well, this presence can manifest in the form of thoughtful representations that provide a felt sense of the human community that preceded us, support the loving bonds that do not disappear with grief, spawn new and richer mourning practices, prompt reflective engagement with our own mortality and those we have lost forever, and contextualize our lives within a chain of being much larger than our selves. In this respect, chatbots can be described as a technology of transcendence. The kind of transcendence at issue here is not the one envisioned by transhumanists (Bostrom 2005). It is not an escape from the human condition. On the contrary, it is a transcendence that is central to the human condition, the sort that is characteristic of, and even necessary for, living well as reflective mortal beings who are capable of attachment, loss, imagination, and remembrance. As we envision them, chatbots of the dead exist within the universal register of art, imagination, and spirit. This register plays an ineliminable role in human flourishing, and we are optimistic that chatbot “actors” can be conducive to a form of flourishing peculiarly suited to our age.²⁷

This technology is still highly novel. Many questions about it remain, both within and beside our framework. These unsettled questions contribute to unease and the sense of danger that many commentators seem to feel about it. More work is needed to understand these types of representations. If this work is to be innovative and expansive—as is necessary to avoid the mistakes that stem from a limited vision of what this technology might be—then this work must be creative and interdisciplinary, open to insights not only from philosophy and computer science but also art, literature, religion, psychology, and other diverse sources of wisdom. If nothing else, we hope to have demonstrated this methodological approach here.

Acknowledgements

We owe thanks to Juls Altman, Caro Flores, Eleanor Helms, Ryan Jenkins, Patrick Lin, Sarah Mantell, Cameron McKean, Paul Miklowitz, Allison Myers, Anna Puzio, Nellie Wieland, Sam Zahn, two anonymous reviewers for this journal, and the audiences at the 2024 Philosophy of Avatars Workshop at the University of Aberdeen, the 2024 Student Symposium at California Polytechnic State University in San Luis Obispo, and a 2024 colloquium talk in the philosophy department at the University of California, Santa Cruz. We owe special thanks to Catelynn Kenner and Jacob Sparks, who generously talked a great deal with us about this project.

References

Adamopoulou, Eleni and Lefteris Moussiades (2020). An Overview of Chatbot Technology. In Ilias Maglogiannis, Lazaros Iliadis, and Elias Pimenidis (Eds.), Artificial Intelligence Applications and Innovations (373–383). Springer. https://doi.org/10.1007/978-3-030-49186-4_31

Adedeji, Joel A. (1972). The Origin and Form of the Yoruba Masque Theatre. Cahiers d’Études Africaines, 12(46), 254–276.

Aristotle (c. 335 BCE/1941). Poetics. In Richard McKeon (Ed.) and Ingram Bywater (Trans.), The Basic Works of Aristotle (1453–1487). Random House.

Averill, James R. (1968). Grief: Its Nature and Significance. Psychological Bulletin, 70(6, Pt. 1), 721–748. https://doi.org/10.1037/h0026824

Bao, Aorigele and Yi Zeng (2024). Embracing Grief in the Age of Deathbots: A Temporary Tool, Not a Permanent Solution. Ethics and Information Technology, 26(1), 7. https://doi.org/10.1007/s10676-024-09744-y

Barish, Jonas A. (1966). The Antitheatrical Prejudice. Critical Quarterly, 8(4), 329–348. https://doi.org/10.1111/j.1467-8705.1966.tb01315.x

Beck, Lauren R. (2024). The Telelibrary Conspiracy: An Autonomous Audience & Their Creation of Remote Community. In Carolyn Ownbey and Catherine Quirk (Eds.), Pandemic Play (185–196). Palgrave Macmillan. https://doi.org/10.1007/978-3-031-47312-8_13

Bennett, Michael Y. (2024). Between the Lines: A Philosophy of Theatre. Oxford University Press.

Blocker, H. Gene (1977). A New Look at Aesthetic Distance. The British Journal of Aesthetics, 17(3), 219–229.

Boal, Augusto (1985). Theatre of the Oppressed (Charles A. McBride and Maria-Odilia Leal McBride, Trans.). Theatre Communications Group.

Bos, Edwin (1995). Making the Dead Live On: An Interactive, Talking Picture of a Deceased Person. ACM SIGCAS Computers and Society, 25(1), 7–9. https://doi.org/10.1145/202481.202485

Bostrom, Nick (2003). Ethical Issues in Advanced Artificial Intelligence. In Iva Smit, Wendell Wallach, and George Eric Lasker (Eds.), Cognitive, Emotive and Ethical Aspects of Decision Making in Humans and in Artificial Intelligence (Vol. 2, 12–17). International Institute for Advanced Studies in Systems Research and Cybernetics.

Bostrom, Nick (2005). Transhumanist Values. Journal of Philosophical Research, 30(Supplement), 3–14. https://doi.org/10.5840/jpr_2005_26

Bowlby, John and Colin Murray Parkes (1970). Separation and Loss within the Family. In Elwyn James Anthony and Cyrille Koupernik (Eds.), The Child in His Family (Vol. 1, 197–216). John Wiley & Sons.

Bradford, Gwen (2013). The Value of Achievements. Pacific Philosophical Quarterly, 94(2), 204–224. https://doi.org/10.1111/j.1468-0114.2012.01452.x

Bradley, Ben (2009). Well-Being and Death. Oxford University Press.

Brake, Elizabeth (2012). Minimizing Marriage: Marriage, Morality, and the Law. Oxford University Press.

Brecht, Bertolt (2002). Short Description of a New Technique in Acting which Produces an Alienation Effect. In Teresa Brayshaw and Noel Witts (Eds.), The Twentieth-Century Performance Reader (2nd ed.). Routledge.

Brooker, Charlie (Writer), & Harris, Owen (Director). (2013, February 11). Be Right Back. In Black Mirror (Season 2, Episode 1).

Bryson, Joanna (2010). Robots Should Be Slaves. In Yorick Wilks (Ed.), Close Engagements with Artificial Companions: Key Social, Psychological, Ethical and Design Issues (63–74). John Benjamins Publishing Company.

Buben, Adam (2015). Technology of the Dead: Objects of Loving Remembrance or Replaceable Resources? Philosophical Papers, 44(1), 15–37. https://doi.org/10.1080/05568641.2015.1014538

Carroll, Noël (1986). Art and Interaction. The Journal of Aesthetics and Art Criticism, 45(1), 57–68. https://doi.org/10.2307/430466

Chalmers, David J. (2010). The Singularity: A Philosophical Analysis. Journal of Consciousness Studies, 17(9–10), 7–65.

Church, Scott H. (2013). Digital Gravescapes: Digital Memorializing on Facebook. The Information Society, 29(3), 184–189. https://doi.org/10.1080/01972243.2013.777309

Csikszentmihalyi, Mihaly (1997). Finding Flow: The Psychology of Engagement with Everyday Life (1st ed.). Basic Books.

Danto, Arthur C. (1974). The Transfiguration of the Commonplace. The Journal of Aesthetics and Art Criticism, 33(2), 139–148.

De Palma, Brian, Robert Fiore, and Bruce Rubin (Directors) (1970). Dionysus in 69 [Film]. https://hdl.handle.net/2333.1/mcvdncsq

del Caso, Borrell (1874). Escapando de la Crítica [Oil on canvas]. Banco de España, Madrid.

Delistraty, Cody (2024). The Grief Cure: Looking for the End of Loss. HarperCollins.

Despret, Vinciane (2021). Our Grateful Dead: Stories of Those Left Behind (Stephen Muecke, Trans.). University of Minnesota Press.

Dimensions in Testimony (2025). USC Shoah Foundation. Retrieved 2 April 2025, from https://sfi.usc.edu/dit

Durkheim, Émile (1912/2008). The Elementary Forms of Religious Life (Carol Cosman, Trans.). Oxford University Press.

Elder, Alexis (2020). Conversation From Beyond the Grave? A Neo-Confucian Ethics of Chatbots of the Dead. Journal of Applied Philosophy, 37(1), 73–88. https://doi.org/10.1111/japp.12369

Epicurus (c. 300 BCE/2000). Letter to Menoeceus (Robert Drew Hicks, Trans.). http://classics.mit.edu/Epicurus/menoec.html

Euripides (2002). Bacchae. In David Kovacs (Ed. & Trans.), Bacchae ; Iphigenia at Aulis ; Rhesus. Harvard University Press.

Fabry, Regina E. and Mark Alfano (2024). The Affective Scaffolding of Grief in the Digital Age: The Case of Deathbots. Topoi, 43(3), 603–614. https://doi.org/10.1007/s11245-023-09995-2

Fagone, Jason (2021). The Jessica Simulation: Love and Loss in the Age of A.I. The San Francisco Chronicle. https://www.sfchronicle.com/projects/2021/jessica-simulation-artificial-intelligence/

Farman, Abou (2020). On Not Dying: Secular Immortality in the Age of Technoscience. University of Minnesota Press.

Feagin, Susan L. (1988). Imagining Emotions and Appreciating Fiction. Canadian Journal of Philosophy, 18(3), 485–500. https://doi.org/10.1080/00455091.1988.10717187

Feldman, Fred (1991). Some Puzzles About the Evil of Death. The Philosophical Review, 100(2), 205–227. https://doi.org/10.2307/2185300

Figueiredo, Madeline de (2024, March 22). Our Last, Impossible Conversation. The New York Times. https://www.nytimes.com/2024/03/22/style/modern-love-ai-our-last-impossible-conversation.html

Fischer-Lichte, Erika (2008). The Transformative Power of Performance: A New Aesthetics (Saskya Iris Jain, Trans.). Routledge.

Floridi, Luciano (2015). “The Right to Be Forgotten”: A Philosophical View. Jahrbuch für Recht und Ethik / Annual Review of Law and Ethics, 23(1), 163–179.

Freiman, Ori and Boaz Miller (2020). Can Artificial Entities Assert? In Sanford Goldberg (Ed.), The Oxford Handbook of Assertion (413–434). Oxford University Press. https://doi.org/10.1093/oxfordhb/9780190675233.013.36

Friedman, Cindy (2023). Ethical Concerns with Replacing Human Relations with Humanoid Robots: An Ubuntu Perspective. AI and Ethics, 3(2), 527–538. https://doi.org/10.1007/s43681-022-00186-0

Gibbons, Michelle G. (2025). Rhetorical Ventriloquism: From Cameos to Deepfakes. Rhetoric Review, 44(2), 101–115. https://doi.org/10.1080/07350198.2025.2462400

Gibson, Margaret (2007). Death and mourning in technologically mediated culture. Health Sociology Review, 16(5), 415–424. https://doi.org/10.5172/hesr.2007.16.5.415

Goffman, Erving (1974). Frame Analysis: An Essay on the Organization of Experience. Harvard University Press.

Goldstein, Rebecca Newberger (2014). Plato at the Googleplex: Why Philosophy Won’t Go Away. Atlantic Books.

Halliwell, Stephen (1986). Aristotle’s Poetics. Gerald Duckworth.

Hamilton, James R. (2017). What is the Relationship between “Observed” and “Participatory” Performance? In Tom Stern (Ed.), The Philosophy of Theatre, Drama and Acting 137–164. Rowman & Littlefield International.

Hamilton, James R. (2019). Philosophy of Theater. In Edward N. Zalta (Ed.), The Stanford Encyclopedia of Philosophy (Winter 2019 Edition). https://plato.stanford.edu/archives/win2019/entries/theater/

Hershenov, David B. (2007). A More Palatable Epicureanism. American Philosophical Quarterly, 44(2), 171–180.

Hollanek, Tomasz and Katarzyna Nowaczyk-Basińska (2024). Griefbots, Deadbots, Postmortem Avatars: On Responsible Applications of Generative AI in the Digital Afterlife Industry. Philosophy & Technology, 37(2), 63. https://doi.org/10.1007/s13347-024-00744-w

Johansson, Jens (2013). The Timing Problem. In Ben Bradley, Fred Feldman, & Jens Johansson (Eds.), The Oxford Handbook of Philosophy of Death (255–273). Oxford University Press.

Karr, Mary (2015). The Art of Memoir. HarperCollins.

Keller, John A. (2015). Paraphrase, Semantics, and Ontology. In Dean W. Zimmerman and Karen Bennett (Eds.), Oxford Studies in Metaphysics (Vol. 9, 90–128). Oxford University Press.

Klass, Dennis (1996). Grief in an Eastern Culture: Japanese Ancestor Worship. In Dennis Klass, Phyllis R. Silverman, and Steven L. Nickman (Eds.), Continuing Bonds: New Understandings of Grief (59–70). Routledge.

Klass, Dennis, Phyllis R. Silverman, and Steven L. Nickman (Eds.) (1996). Continuing Bonds: New Understandings of Grief. Routledge.

Kroon, Fred and Alberto Voltolini (2019). Fiction. In Edward N. Zalta (Ed.), The Stanford Encyclopedia of Philosophy (Winter 2019 Edition). https://plato.stanford.edu/archives/win2019/entries/fiction/

Krueger, Joel and Lucy Osler (2019). Engineering Affect: Emotion Regulation, the Internet, and the Techno-Social Niche. Philosophical Topics, 47(2), 205–232.

Krueger, Joel and Lucy Osler (2022). Communing with the Dead Online: Chatbots, Grief, and Continuing Bonds. Journal of Consciousness Studies, 29(9–10), 222–252. https://doi.org/10.53765/20512201.29.9.222

Kurzweil, Amy (2023). Artificial: A Love Story. Catapult.

Kurzweil, Ray (2005). The Singularity Is Near: When Humans Transcend Biology. Penguin.

Kwastek, Katja (2013). Aesthetics of Interaction in Digital Art (Niamh Warde, Trans.). MIT Press.

Levy, David (2008). Love and Sex with Robots: The Evolution of Human-Robot Relations. Harper Perennial.

Lindemann, Nora Freya (2022). The Ethics of ‘Deathbots’. Science and Engineering Ethics, 28(6), 1–15. https://doi.org/10.1007/s11948-022-00417-x

Lingel, Jessa (2013). The Digital Remains: Social Media and Practices of Online Grief. The Information Society, 29(3), 190–195. https://doi.org/10.1080/01972243.2013.777311

Liu, Pengbo (2025). Preservation or Transformation: A Daoist Guide to Griefbots. In Philipp Hacker (Ed.), Oxford Intersections: AI in Society. Oxford University Press. https://doi.org/10.1093/9780198945215.003.0067

Locker, Melissa (2015, October 8). Warm-Up Comics: The Funniest People in TV Who Are Never on TV. Vanity Fair. Retrieved 16 August 2024, from https://www.vanityfair.com/hollywood/2015/10/tv-warm-up-comics

Lopes, Dominic (2009). A Philosophy of Computer Art. Routledge.

Mallory, Fintan (2023). Fictionalism about Chatbots. Ergo, 10(38). https://doi.org/10.3998/ergo.4668

Mandelbaum, David G. (1959). Social Uses of Funeral Rites. In Herman Feifel (Ed.), The Meaning of Death (189–217). McGraw-Hill.

Marsh, Walter (2024, February 28). Laurie Anderson on Making an AI Chatbot of Lou Reed: ‘I’m Totally, 100%, Sadly Addicted’. The Guardian. https://www.theguardian.com/music/2024/feb/28/laurie-anderson-ai-chatbot-lou-reed-ill-be-your-mirror-exhibition-adelaide-festival

McMahan, Jeff (1988). Death and the Value of Life. Ethics, 99(1), 32–61.

McTaggart, John McTaggart Ellis (1908). The Unreality of Time. Mind, 17(68), 457–474.

Meese, James, Bjorn Nansen, Tamara Kohn, Michael Arnold, and Martin Gibbs (2015). Posthumous Personhood and the Affordances of Digital Media. Mortality, 20(4), 408–420. https://doi.org/10.1080/13576275.2015.1083724

Miklowitz, Paul S. (2024). Our Minds, Our Selves: Minds, Meaning, and Machines. Borderless Philosophy, 7, 109–131.

Moss, Miriam S. and Sidney Z. Moss (1996). Remarriage of Widowed Persons: A Triadic Relationship. In Dennis Klass, Phyllis R. Silverman, and Steven L. Nickman (Eds.), Continuing Bonds: New Understandings of Grief (163–178). Routledge.

Nabokov, Vladimir (1999). Speak, Memory: An Autobiography Revisited. Alfred A. Knopf.

Newton, Casey (2016, October 6). Speak, Memory: When Her Best Friend Died, She Rebuilt Him Using Artificial Intelligence. The Verge. Retrieved from https://www.theverge.com/a/luka-artificial-intelligence-memorial-roman-mazurenko-bot

Nguyen, C. Thi (2020). Games: Agency as Art. Oxford University Press.

Nguyen, C. Thi (2021). How Twitter Gamifies Communication. In Jennifer Lackey (Ed.), Applied Epistemology (410–436). Oxford University Press.

Nissenbaum, Helen (2004). Privacy as Contextual Integrity. Washington Law Review, 79(1), 119–158.

Norlock, Kathryn J. (2017). Real (and) Imaginal Relationships with the Dead. The Journal of Value Inquiry, 51(2), 341–356. https://doi.org/10.1007/s10790-016-9573-6

Öhman, Carl and Luciano Floridi (2017). The Political Economy of Death in the Age of Information: A Critical Approach to the Digital Afterlife Industry. Minds and Machines, 27(4), 639–662. https://doi.org/10.1007/s11023-017-9445-2

Öhman, Carl and Luciano Floridi (2018). An Ethical Framework for the Digital Afterlife Industry. Nature Human Behaviour, 2(5), 318–320. https://doi.org/10.1038/s41562-018-0335-2

Puzio, Anna (2023). When the Digital Continues After Death: Ethical Perspectives on Death Tech and the Digital Afterlife. Communicatio Socialis, 56(3), 427–436. https://doi.org/10.5771/0010-3497-2023-3-427

Quine, Willard Van Orman (1948). On What There Is. The Review of Metaphysics, 2(5), 21–38.

Rachels, James (1975). Why Privacy is Important. Philosophy & Public Affairs, 4(4), 323–333.

Rodogno, Raffaele (2016). Social Robots, Fiction, and Sentimentality. Ethics and Information Technology, 18(4), 257–268. https://doi.org/10.1007/s10676-015-9371-z

Rosenberg, Harold (1952). The American Action Painters. Art News, 51(8), 22–23, 48–50.

Rosenblatt, Paul C. and Cynthia Meyer (1986). Imagined Interactions and the Family. Family Relations, 35(2), 319–324. https://doi.org/10.2307/583641

Rosenblatt, Paul C., R. Patricia Walsh, and Douglas A. Jackson (1976). Grief and Mourning in Cross-Cultural Perspective. Human Relations Area Files.

Rothaupt, Jeanne W. and Kent Becker (2007). A Literature Review of Western Bereavement Theory: From Decathecting to Continuing Bonds. The Family Journal, 15(1), 6–15. https://doi.org/10.1177/1066480706294031

Ryle, Gilbert (1949). The Concept of Mind. University of Chicago Press.

Sajadieh, Sahar and Hannen Wolfe (2023). Designing Interrogative Robot Theater: A Robot Who Won’t Take No for an Answer. In Anthony L. Brooks (Ed.), Creating Digitally (183–211). Springer.

Salen, Katie and Eric Zimmerman (2004). Rules of Play: Game Design Fundamentals. MIT Press.

Saltz, David Z. (1997). The Art of Interaction: Interactivity, Performativity, and Computers. The Journal of Aesthetics and Art Criticism, 55(2), 117–127. https://doi.org/10.2307/431258

Schechner, Richard (1994). Environmental Theater. Applause Books.

Schechtman, Marya (2025). Talking to Myself: AI and Self-Knowledge. Social Epistemology, 39(4), 1–10. https://doi.org/10.1080/02691728.2025.2480274

Scheutz, Matthias (2011). The Inherent Dangers of Unidirectional Emotional Bonds between Humans and Social Robots. In George A. Bekey, Keith Abney, and Patrick Lin (Eds.), Robot Ethics: The Ethical and Social Implications of Robotics (205–221). MIT Press.

Sharkey, Amanda and Noel Sharkey (2012). Granny and the Robots: Ethical Issues in Robot Care for the Elderly. Ethics and Information Technology, 14(1), 27–40. https://doi.org/10.1007/s10676-010-9234-6

Shields, Christopher (2023). Aristotle. In Edward N. Zalta and Uri Nodelman (Eds.), The Stanford Encyclopedia of Philosophy (Winter 2023 Edition). https://plato.stanford.edu/archives/win2023/entries/aristotle/

Silverman, Phyllis R. and Dennis Klass (1996). Introduction: What’s the Problem? In Dennis Klass, Phyllis R. Silverman, and Steven L. Nickman (Eds.), Continuing Bonds: New Understandings of Grief (3–27). Routledge.

Silverstein, Harry S. (1980). The Evil of Death. The Journal of Philosophy, 77(7), 401–424. https://doi.org/10.2307/2025537

Somers, Meredith (2020, July 21). Deepfakes, Explained. MIT Sloan School. Retrieved from https://mitsloan.mit.edu/ideas-made-to-matter/deepfakes-explained

Sparrow, Robert and Linda Sparrow. (2006). In the hands of machines? The future of aged care. Minds and Machines, 16(2), 141–161. https://doi.org/10.1007/s11023-006-9030-6

Stokes, Patrick (2012). Ghosts in the Machine: Do the Dead Live on in Facebook? Philosophy & Technology, 25(3), 363–379. https://doi.org/10.1007/s13347-011-0050-7

Stokes, Patrick (2015). Deletion as Second Death: The Moral Status of Digital Remains. Ethics and Information Technology, 17(4), 237–248. https://doi.org/10.1007/s10676-015-9379-4

Stokes, Patrick (2021). Digital Souls: A Philosophy of Online Death. Bloomsbury Academic.

Story, Daniel and Ryan Jenkins (2023). Deepfake Pornography and the Ethics of Non-Veridical Representations. Philosophy & Technology, 36(3), 56. https://doi.org/10.1007/s13347-023-00657-0

Stroebe, Margaret, Mary Gergen, Kenneth Gergen, and Wolfgang Stroebe (1996). Broken Hearts or Broken Bonds? In Dennis Klass, Phyllis R. Silverman, and Steven L. Nickman (Eds.), Continuing Bonds: New Understandings of Grief (31–44). Routledge.

Sweeney, Paula (2021). A Fictional Dualism Model of Social Robots. Ethics and Information Technology, 23(3), 465–472. https://doi.org/10.1007/s10676-021-09589-9

Sweeney, Paula (2024). Social Robots: A Fictional Dualism Model. Rowman & Littlefield.

Timmerman, Travis (2022). Dissolving Death’s Time-of-Harm Problem. Australasian Journal of Philosophy, 100(2), 405–418. https://doi.org/10.1080/00048402.2021.1891108

Undersigned (2022, August 8). Everything Immersive. Retrieved 7 April, 2025, from https://everythingimmersive.com

Vallor, Shannon (2024). The AI Mirror: How to Reclaim Our Humanity in an Age of Machine Thinking. Oxford University Press.

van den Hoven, Jeroen (2008). Information Technology, Privacy, and the Protection of Personal Data. In Jeroen van den Hoven and John Weckert (Eds.), Information Technology and Moral Philosophy (301–321). Cambridge University Press.

von Solodkoff, Tatjana (2014). Paraphrase Strategies in Metaphysics. Philosophy Compass, 9(8), 570–582. https://doi.org/10.1111/phc3.12150

Walton, Kendall L. (1990). Mimesis as Make-Believe: On the Foundations of the Representational Arts. Harvard University Press.

Warren, James (2004). Facing Death: Epicurus and His Critics. Oxford University Press.

Weber-Guskar, Eva (2021). How to Feel About Emotionalized Artificial Intelligence? When Robot Pets, Holograms, and Chatbots Become Affective Partners. Ethics and Information Technology, 23(4), 601–610. https://doi.org/10.1007/s10676-021-09598-8

Weizenbaum, Joseph (1966). ELIZA—A Computer Program for the Study of Natural Language Communication between Man and Machine. Communications of the ACM, 9(1), 36–45. https://doi.org/10.1145/365153.365168

Wright, Nicola (2014). Death and the Internet: The Implications of the Digital Afterlife. First Monday, 19(6). https://doi.org/10.5210/fm.v19i6.4998