Digital Souls: A Philosophy of Online Death
by Patrick Stokes
Published January 2021
On 23 June, the American entrepreneur John McAfee died in a Spanish prison. The details of his death are the subject of much conspiratorial theorizing, spurred on by McAfee’s lifestyle and self-consciously cultivated legend. Soon after his death, an image of the letter Q was posted on his Instagram account, in a nod to the far-right QAnon conspiracy theory. It’s more than likely that McAfee, being the shitposter that he was in life, simply instructed someone on his staff to do so in the event of his death. But what if the explanation wasn’t so simple?
Patrick Stokes’ latest book, Digital Souls: A Philosophy of Online Death, begins with a similar, but more unnerving example. Tim is sitting on his couch one evening when he receives an email with the subject ‘I’m watching.’ The body of the email was to the point as well: ‘Did you hear me? I’m at your house. Clean your fucking attic!!! – Jack Fröse.’ While a message like this from a stranger would be disturbing, Tim and Jack were lifelong friends. Jack, however, had passed away a few months earlier. His message seemed to have been sent from beyond the grave. Tim wasn’t the only one of Jack’s friends to receive a message, and though he didn’t reply, those who did never received a response.
We could put the message from Jack down to a glitch in cyberspace. A simple explanation would be that it just got held up and arrived months after actually being sent, giving Tim a fright. Or perhaps that it was a morbid prank to get a rise out of Jack’s friends and relatives? According to Jack’s mother, that didn’t matter. For her, it was as if the messages were from Jack, and what mattered is that they helped everyone remember him. In some respects, these messages from the afterlife were a clear sign of Jack’s presence to his relatives and friends, and they suggested that perhaps, in some way he had persisted past his death in the digital world.
For Stokes, the most compelling aspect of Tim and Jack’s story however lies in the insight it gives us into our relationship with the dead and with technology. It is this relationship that is the focal point of Digital Souls. In this book, Stokes shows us that while there is no shortage of discourse about the different ways in which our lives are affected by the internet, we must equally consider how our deaths and our relationship to death and the dead are changing through our use of technology, particularly social media.
The worry is this: as we live more of our lives digitally, in spaces facilitated by the internet, social media, and our devices, we leave behind an increasing volume of what Stokes calls ‘digital remains’: social media profiles, email archives, hard drives, photos, message logs, forum posts, blogs and so on. All these things exist at the intersection of a number of philosophical problems related to questions about identity, personhood, the nature of grief, our obligations towards the dead, and the shift in our moral universe caused by new technologies. For Stokes, it is not simply the case that modern technology mediates our relationship to death in an unpredictable and under-theorised way, though he does think it is doing both of those things. As he reminds us, technology has always affected the way we thought of our relationship to the dead, and it has always fueled our imagination and shifted our conceptions of dying. Rather, as technology dominates more of our lives, a significant part of being human happens online. And if to live as a human being is to live digitally, then so it is to die. The problem, Stokes argues, is that it’s not entirely clear in what way we die as far as our digital self is concerned.
What is the ontological status of our digital persona? One answer is that it has no status. While the old meme might have it that on the internet nobody knows you’re a dog, we all know that behind the screen, behind the avatars we interact with daily, there are real people, made of meat and bones, who inhabit a digital space for a time, and who can log off anytime they want. To Stokes, this isn’t a very satisfying solution because it doesn’t seem to capture the nature of our digital presence. To him, we aren’t solely individuals hiding behind our avatars, in an important sense we are our avatars.
He gets to this idea via the extended cognition thesis, the view that our cognitive capacities can be outsourced to various forms of technology, and not merely contained in our bodies and minds. Consider, for example, how few phone numbers most of us remember these days – this is because our phones are doing it for us, thus extending our mental capacity. The extended mind thesis takes this view one step further. This is the idea that it is not just our intellectual powers that can be extended in this way, but that the mind itself too can inhabit, at least partly, the external world. To illustrate, Stokes cites the philosopher Richard Heersmirk, who writes: ‘The complex web of relations we develop and maintain with other people and artefacts partly constitutes our self, implying we are essentially a soft self. Personal identity can thus neither be reduced to psychological structures instantiated by the brain nor by biological structures instantiated by our biological organism.’ The term ‘soft self’ is itself taken from philosopher Andy Clark, who defines it as ‘a constantly negotiable collection of resources easily able to straddle and criss-cross the boundaries between biology and artefact.’
Stokes takes these ideas as the starting point for building his argument that our digital life needs to be considered after we die. He argues, we exist not simply as bodies, ‘but “smeared” across the world through all the things we bring into our personhood and through which we are instantiated.’ That is, as a ‘soft self.’ To him, our digital footprint is not merely a collection of data located in a server farm somewhere in the middle of nowhere. Rather, it is a part of our self that contains an essential part of our personhood. The things we leave behind, our ‘digital remains,’ are therefore aspects of us that can persist beyond our death. Though the individual might cease to exist in the meatspace, parts of their extended, soft self remain intact in cyberspace.
The key to this is the way in which social media and other online activities give others a sense of our presence. Stokes reaches to Walter Benjamin to make his case. Benjamin famously argues that the mechanical reproduction of art robs it of its ‘aura.’ A poster from a museum giftshop, after all, is not the masterpiece in the gallery. At the same time, Benjamin makes an exception for photographs, notably for their ability to capture the aura of the dead. He writes:
The cult of remembrance of loved ones, absent or dead, offers a last refuge for the cult value of the picture. For the last time the aura emanates from the early photographs in the fleeting expression of the human face.
The same can be said of our digital presence, at least in some relevant instances. We have always left behind various traces of our existence. A brooch that belonged to your grandmother might be priceless to you, and its loss would be immeasurable to you, even if it would be be worthless to a collector. What distinguishes our digital lives, and particularly social media is that, in Stokes’ words, ‘they are born as vehicles for identity performance and maintenance. It is built to present you specifically and individually; it gives out your identity as part of its essential construction.’ Social media is designed to capture you.
If we take seriously the idea that our personhood in some way inhabits cyberspace, what should we do with the digital remains we leave behind? If, as Stokes argues, these things are our aura, then it would be wrong to simply delete this data. On this view, to shut down a dead relative’s Facebook profile would not merely be to strike her off the corporate behemoth’s database, but to delete a part of her that is still real, and to do so might be doing her a real moral harm. Moreover, it might also do harm to the living, by depriving them of an important and meaningful way of grieving and interacting with the dead.
Stokes is right to point to these problems. Consider what a social media profile is. In a basic, perhaps idealised version, it’s a place for a person to share photos and communicate with their friends. Those of us who remember Myspace, and who learned HTML and CSS thanks to it, will also understand that a social media profile is a way of presenting yourself to the world that you have some control over. That is, it is a curated version of the self that we present to others. For example, though all Facebook profiles look more or less the same in terms of the basic layout, the content is all chosen by the users and put together it amounts to a version of themselves that they’d like to project into the world. This is equally true of other platforms, like Twitter, LinkedIn, and even the far-right Parler. Sometimes even belonging to the space is intended to give others some impression about us. These are spaces where we can shape the narratives that define our lives. As Stokes is quick to point out, our online and ‘real’ narratives intersect and influence each other. When we go online, we never truly abandon the physical world. For instance, facial recognition algorithms online can be used to precisely match our digital and meat selves, while the biometric tracking we all increasingly subject our bodies to via our mobile phones and fitness gadgets can update our social media in real-time. Certain things we do online will have clear, real-world consequences (consider people who posted something awful on Twitter and lost their jobs as a result), and vice versa.
At the same time, social media is not perfectly curated – as our conversations happen not just in the publicly visible portions of our profiles but in private messages and group chats, there is also scope for being spontaneous and uncurated on the internet. Though Stokes doesn’t touch on this directly, he does make it clear that we need to approach these questions holistically. This leaves open the question of what digital remains are most valuable. Social media is the simplest example Stokes addresses, perhaps because it feels most familiar – but I take his point to be broader. We also leave behind emails, photo archives, medical records, GPS tracking data, and other less obvious detritus. If we focus too narrowly on which of the artefacts we leave behind are valuable, we would lose the multifaceted way in which we interact with others online. We present different versions of ourselves to different communities and people, and though they are all version of us, it’s not impossible that to our close ones they would be completely unrecognisable. And though I can’t imagine many people would feel nostalgic for a deceased relative’s medical records (though they may be valuable in other ways), one can easily see how, for example, someone’s private correspondence can evoke the kind of aura Stokes is talking about. Even GPS tracking data could lead one on an adventure retracing a loved relative’s steps.
The less our digital presence can be separated from the rest of our lives, the more it becomes clear that our identity and personhood are tied up with what we do online. Our avatar is often the version of ourselves that some of our friends interact with the most, all the more so as we live through a global pandemic. When we die, then, it is a natural extension of the way we live that our digital remains can become a locus for other people’s grief. It’s not uncommon, as Stokes shows, that a Facebook profile becomes a space where friends and relatives can commemorate the person they lost, by posting tributes or remembrances. This practice has become so common that in some instances it is now possible for the relatives of a dead person to convert their page into a memorial, and to gain a level of control over it. This, however, leads to another problem. If a social media profile captures our personhood and allows it to persist after death, we are at risk of becoming overwritten by the tributes left behind by mourners. As more and more loved ones leave messages, less and less of the profile is created by the dead person. As Stokes writes,
Like pilgrims touching or kissing a relic or statue, very gradually wearing it away through their gentle, loving contact, every interaction with an online profile of the dead may imperceptibly contribute to destroying what it pays tribute to.
Preservation also offers an additional challenge. Eventually, social media sites and their servers will become overrun with the profiles of dead people. It is easy enough to expect these companies to maintain these profiles while the amount of them is trivial with regards to their total user-bases. For now, keeping these profiles online provides these companies a financial incentive. Dead individuals might not generate profit directly – they aren’t likely to click on ads – but those they leave behind still offer live eyeballs and engagement when they interact with their loved ones. Stokes quips, following Marx, that there has never been a better illustration of ‘dead labor.’ But this is likely enabled by the relatively small cost of keeping a minority of accounts online. To go along with Stokes’ argument, at some point in time, the living might become a minority in these spaces, turning them into online cemeteries. What then?
As is often the case, the solution to a problem posed by new technologies isn’t likely to be simply moving away from it, but rather the emergence of a newer technology. Stokes gives us a few possible outcomes. One example is that of Roman Mazurenko, a Russian entrepreneur who in 2015 tragically died in a car accident. One of his friends, a co-founder of an artificial intelligence start-up, took the initiative to scrape Mazurenko’s digital remains to create a chatbot, which would enable any of his friends to chat to him, or at least, to something that resembled him in his speech patterns and mannerisms. Those of his friends who were willing to try it say the result is chillingly like speaking with Mazurenko himself. And this provoked a mixed reaction. Some found it comforting to be able to unload their grief into a chatbot – they felt it allowed them to say what they didn’t get a chance to while their friend was still alive. For others, it was disturbing or even distressing to interact with Mazurenko in that way. Instead of allowing them to move on, the chatbot suspended their grief. This is perhaps easy enough to avoid. Knowing what the bot is, or what it purports to be, one can simply avoid interacting with it. Stokes points to a bigger, moral issue. If our digital remains already allow us to persist beyond death in a meaningful way, then the living acquire moral obligations in relation to them. This is to say, it might be wrong to use someone’s digital remains in the way that Mazurenko’s friend did.
This intuition seems to be broadly shared, as evidenced by the recent backlash received by Morgan Neville, the director of the documentary Roadrunner: A Film About Anthony Bourdain, which uses AI to effectively reanimate the late celebrity chef to narrate the story. Though the film’s use of digital recreation to voice Bourdain is thin (reportedly there are only three instances of this being done in the film), it nonetheless carries a certain moral stench that many of Bourdain’s fans immediately noticed. Besides the ethical issues with the living using the dead in this way, we should also worry about the accuracy with which digital means can bring someone back; after all, our digital presence tends to be curated, even if imperfectly. As such, at best it seems we can bring back an idealized approximation of the person.
It is examples like these gives that I find most worrying. If, as Stokes argues, we survive via our digital remains, can we be meaningfully brought back to life? A chatbot like that made by Mazurenko’s friend is not likely to convince anyone that the man was still alive. Companies that promise this sort of reanimation, Stokes notes, tend to be quite modest in their claims. They might claim to preserve you in some sense, they don’t claim that their offering would become you after you die. That is to say, one could feed all of a loved one’s digital life to a neural network to create such a chatbot, and nobody would be under the illusion that the chatbot is the person in question. What changes after one’s death is that the chatbot might simply be the closest thing to the real person, which can give us an adequate illusion of their presence. At the same time, Stokes reminds us, the study of consciousness is far away from solving the so-called ‘hard problem’ of consciousness. That is, the problem of understanding how our subjective personal experiences come about from the biological processes occurring in our nervous systems (for the sake of argument here, let’s suppose that the biological processes are all there is to our consciousness). If we can’t understand how these processes occur in general (yet), then we equally can’t reproduce them in a computer simulation. This is to say, the bot would not be the actual person – it wouldn’t have their inner life, history, thoughts and feelings. At best, it would be an approximation.
But these limitations could be overcome with time, and such emulations of life could become good enough to genuinely fool us. Then, the question of how we interact with the dead – and the ends to which we bring them back – acquires some urgency. To some the pain of losing loved ones might make the possibility of bringing them back in a digital form attractive, especially if the simulation is convincing enough that one could forget that it isn’t real. Stokes illustrates this with an example from the British TV series Black Mirror. In one episode a woman grieving her husband acquires a digital simulation of him which can be installed in an artificial body resembling his. Although she finds the outcome disturbing, the artificial version of her husband is so convincing that she cannot bring herself to deactivate it when she becomes unhappy with it. Another example, which is perhaps more nefarious, is the way in which deceased actors are brought back through digital means to reprise their famous roles, as was the case in recent Star Wars films, with Peter Cushing and Carrie Fisher. Stokes also points to other examples of dead celebrities – actors and musicians – being similarly reanimated. While bringing a loved one back due to grief might be problematic (to say the least) at least it doesn’t feel outright exploitative in the way bringing dead actors back to life does.
For Stokes, there is a real danger that as these technologies become better, we will lose sight of the distinction between the real and the artificial. And while Roman Mazurenko chatbot might indicate where we might be headed, they also show that there is a risk the technological mediation of our interactions with the dead can retreat into the background or disappear completely. In that case, Stokes warns us that we should be careful of what that will mean to our understanding of ourselves. I worry too that even death doesn’t allow a person to escape from exploitation for profit.
Though it is clear that this book is published on the cusp of what will likely be a vast philosophical literature on dying in the digital age, Stokes nonetheless proposes a solution to some of the issues he brings up. Broadly speaking, he urges caution. Tech companies, he points out, rarely consult with philosophers at the development stage of new products. This is a shame, because as this book shows, philosophers are well-equipped to think through the wide implications new technologies pose for our self-understanding. And while the technologies related to death in the digital space are still in their infancy, we need to think about their implications now. As these things improve, as the bots that can imitate our beloved dead become more realistic, we will, Stokes argues, need a failsafe to remind us of what is reality and what is artificial. He explains that a glitch built into these systems will ensure that. How? By reminding us that there is a boundary there that we shouldn’t cross: ‘glitches break our absorption and break the spell. Perhaps, then, glitches can stop us becoming so engrossed in talking to an avatar that we slip into treating it as a real person.’
Digital Souls is not just a cautionary tale. As much as Stokes worries about what excesses of technology in relation to death, he also makes it clear that these technologies have opened up positive aspects to our relationship with the dead. Social media can, when approached cautiously, bring us closer while we’re alive and enable us to feel close to our beloved dead. Stokes’ writing brings complex and difficult philosophical concepts into the sphere of our everyday lives with ease and flair. In bringing life to writing about death, he invites us to join him in thinking philosophically about how technology is changing our world.