“It is 2015 and we are closer to launching the Eterni.me avatar that will eventually become your digital alter ego, your immortal bits-and-bytes clone,” reads the optimistic first sentence of a Medium post by entrepreneur Marius Ursache. That post, as well as articles by Fast Company, NBC News, and the New Yorker, are the only remnants of Eterni.me, a company that once boasted 30,000 users-in-waiting. The goal was “digital immortality,” an algorithm that makes communicating with you possible long after you’re dead. Unfortunately, Eterni.me proved to be digitally mortal.
Tech companies have started trying to cash in on dead people relatively recently, but the use of stand-ins and representations of departed loved ones precedes our current technology’s digital flavor. Before Kim Kardashian’s hologrammed father there were funereal masks, commemorative portraits, lockets full of hair. When photography presented the possibility of both replicating reality and freezing it in time, the Victorians didn't hesitate to try it on their dead. They took great lengths to peel back the eyelids and prop up the limp bodies of ill-fated children, to render at least their image immortal.
But as the wheels of technological progress turn, the world of technology-borne reanimation still rests primarily on a gossamer web of metaphors and hypotheticals. Like Eterni.me, the products always look promising—in 2015 there was Elysium, a VR world inhabited by your deceased relatives; in 2019, Ghost, an app that sends texts from the grave; in 2018, Roman, a bot that reproduces the conversational style of a lost friend. Their vocabulary is at once bombastic and familiarly hollow—“AI,” “virtual reality,” “artificial neural network,” “chatbot,” “hologram.” You probably don’t know the mechanical specifics of any of these...but someone does.
They enter competitions, flaunt prototypes, create Instagram accounts, court investors. Tech publications and culture magazines write about them under clickbait headlines, almost always referencing the same Black Mirror episode. Sometimes the companies launch. Sometimes they even have users. More often, though, they disappear into the ether, leaving us to our own grieving devices.
When no investors pick up a given digital reincarnation venture, or when an app shutters and the personal data harvested is inevitably sold off, what remains? Only the irresistibly eerie, deliciously controversial narrative that was spun around them.
“Be Right Back,” the Black Mirror episode so often cited, tells the story of Martha, a widow who begins a relationship with an advanced digital replica of her deceased husband, Ash. The replica begins as a chat bot, then becomes a speech-enabled program, and finally is delivered to her door as an embodied clone. By the time Martha reckons with the fact that this android can’t replace a human, she’s already gone too far in imagining him as one. She orders him to jump off a cliff, but when he starts to obey, she reveals how the real Ash would respond. Immediately internalizing her projection, android Ash starts crying, trembling, pleading. The artifice is so obvious, and at the same time, the emotional effect so convincing, that Martha can neither justify keeping him as a human, nor disposing of him as a machine. She ends up hiding him in her attic.
Eugenia Kuyda, the creator of Roman bot, didn’t wait for journalists to make the connection between a dystopian TV show and her creation. She herself references the Black Mirror episode as an inspiration for her company, which she started to memorialize a friend who died in a car accident. The similarities are obvious: Kuyda was driven by grief, and created the bot by scraping all of her friend’s text messages and emails. But the difference in scale is significant. The bot was used less as a stand-in for its original referent, and more as a diary where people dumped their existential woes. Kuyda noticed, and rebranded the program to a virtual “companion” called Replika. It no longer pretends to be someone else—instead, the chatbot digests every text you send it and then spits out the perfect interlocutor: yourself.
That tech advertising and coverage sometimes dips into science fiction is not surprising when so much of flashy technological innovation is about what we think it will be rather than what it is. Elon Musk, who made his fortune off creating a mundane payment app, now spends his days posting about outlandish and improbable inventions like “fitbits in your skull,” much to the delight of his devoted followers. In the case of technology that promises to touch the most intimate parts of our lives, the impulse to compare it to fiction is even stronger.
In Andrei Tarkovsky’s Solaris (1972), the convergence of the scientific and supernatural is conveyed through what is essentially a ghost story. Kris Kelvin, a psychologist, is sent to a space station in the sentient ocean-planet Solaris. Something about the ocean causes the station’s crew members to see the most shameful, repressed, and traumatic depths of their consciousness materialized as physical “guests.” The moment Kris goes to sleep on the station, his memory too is excavated by Solaris, and he wakes up next to his wife, Hari, who killed herself ten years prior.
This Hari is a replica, made up of chemical materials, based entirely on Kris’s perception of her, dependent on his physical presence for survival. Her dress doesn’t have a zipper because it’s only a facsimile, her flesh heals instantly because it isn’t really flesh. At first, Kris tries to get rid of her, but she keeps coming back, and he is powerless against the combination of love, guilt, and nostalgia that connect them.
“Don't turn a scientific problem into a common love story,” one of the more skeptical crew members tells him. In the end, however, Hari’s material makeup and scientific origin are not as important as her relationship with Kris and her personhood. She is not merely a vehicle for contact with the planet, but a being that loves, reflects, and suffers.
Science fiction, as well as the journalism that references it, supposedly upends the supernatural by turning it into the scientific. But the genre lives in the shadow of the mystical. Aren’t all the stories we tell about technology informed by the same spiritual questions, once answered by the paranormal, the magical and otherworldly?
What is the difference between a ghost and a bot? Both are phantoms, but the latter grafts the mysticism at the core of contact with the departed onto our world of skepticism and explanation. “Don’t worry, this isn’t some spiritual thing,” a woman recommending the reanimation bot tells Martha in “Be Right Back.”
The uncanniness of digital resurrection is that it tries to have death both ways—impenetrable and transcendent, as well as mechanistic and decipherable. Of course, no one in the tech world claims that their virtual ghosts are actually communicating from the nether world. And even if they did, machine learning and artificial intelligence technology are nowhere near the competence we see portrayed in science fiction. At the same time, the breathless sensation with which these products are marketed presumes realism, or at least some suspension of disbelief.
Paranormal communications with the dead, on the other hand, require complete naivete. There are no disclaimers or caveats, no “About” pages for reference. And unlike technological appendages, these experiences aren’t always chosen. We can try to explain them away through biology and neuroscience, but the gaping void between that information and our experience will remain. When the specters of our beloved come to us in dreams, we have no choice but to let them.
It was 1977, and a computer technician called John Dilks III had invented a $39,000, bulletproof-glass-encased “talking tombstone.” The attendant People magazine feature opened with a verse: “Here lies John Doe / To hear his tale / Push button below.” Dilks’ headstone was solar-powered, equipped (supposedly) with a mechanical arm that trimmed the surrounding grass. A screen flashed photos of the deceased, read out epigraphs; “imitations of immortality,” People called them.
There is something redundant about mechanizing a tombstone. Graveyards have always functioned as physical, pedestrian archives, “databases of the dead,” as Bjorn Nansen writes, “in which the dead are organized in a managed fashion.” One designer recently drew up a concept of the cemetery as rows of USB memory sticks, slotting into a computer, rather than a grave—a means of digitizing the dead, the designer wrote; a mirage of greater permanence. But USB sticks will erode as tombstones do.
As Dilks’ gravestone demonstrates, immortality is costly. The database of the dead is never complete—to enter, one needs a marked tombstone, a funeral plot, an epigraph. Records are to be bought, and without them, cemeteries are bulldozed away. Even as death, through digital channels, has become ever-present, memorialized around us, on Facebook and in GoFundMe campaigns for funeral costs, the bodies of Covid-19 victims languish in trucks in New York City. They will be buried in the mass grave at Hart Island, because they are left “unclaimed” by the state, or because their families are too poor to buy a grave plot.
Meanwhile, “talking tombstones” continue to proliferate: gravestones of the wealthy are now adorned with QR codes, etched into stone; 48-inch tombstone screens serve as cinematic memorials. Every so often, a new digital gravestone is invented (in 2017, in Slovenia; in the U.S. the next year), and tabloids chorus that it is the “world’s first” such technology, the one that, maybe, will fend off oblivion at last.
what we’re looking at
Occupied Land Ramsey Nasser visualizes the occupation of Palestine