Rohit Prasad, Head Scientist of Alexa, took to the stage at a tech conference in June 2022.
On the giant screen behind him, a young boy asked the virtual assistant if Grandma could read The Wizard of Oz. As her voice issued from the speaker, the boy smiled, with a wistfulness that made it clear his grandmother was no longer alive. Less than a minute of audio sample was required to accomplish this feat.
The presentation spurred several journalists to contact me, asking if I could explain the impact on the bereaved. Wearing the hat of a psychotherapist who often works with loss, I said what I always say: that attempting to stay connected with our dead through currently available technologies is as old as time; that we shouldn’t seek to weigh ourselves or anyone else against mythological grief norms; and that the experience of technologically mediated grieving is different for everyone.
But as a cyberpsychologist who’s studied the intersections of death and the digital for a decade and a half, I’m also interested in the broader phenomenon here: the increasing ease with which the dead can be puppeteered to do our bidding. I’m concerned with the societal, environmental, and economic impacts of deceased people’s data. We could be poised at the dawn of an unprecedented era of dead labour.
By dead labour, Marx meant the machinery, tools, and infrastructure built by previous generations that still underpin the work and profit of the living. In a knowledge economy, the labour of the past is concretised in the digital remains most of us leave behind.
Accessibility and Ease
Not so long ago, you needed big bucks to get the dead to work for you, and the targets of these efforts were mainly focused on celebrities. Director Morgan Neville hired a software company to generate a deepfake of Anthony Bourdain’s voice for a documentary about the late star of television and kitchen. BASE Hologram sent the late Whitney Houston on a three-year world tour. Industrial Light and Magic brought Carrie Fisher and Peter Cushing back to life for cameo roles in 2016’s Rogue One. And who knows how much Kanye West spent on the hologram of Robert Kardashian, the one that wished Kim a happy 40th and did a little dance for her?
But as Prasad demonstrated, reanimating the dead is becoming more advanced, accessible, and affordable. With cheap software, a few scraps of audio or video data, and no special skills, you can produce a convincing digital facsimile of a once-living person. For example, my 12-year-old used my audio clone on Descript to create a series of persuasive-sounding deathbed confessions to play at my funeral. I keep seeing unsettlingly good video chatbots on LinkedIn, purportedly generated in mere moments.
Regulation in this area is so light or non-existent that there’s little to stop people from reanimating the dead for whatever purposes they wish. The famous enjoy increasing protections in some jurisdictions, including America, where the 2023 SAG-AFTRA strike resulted in new protections that severely limit applications of AI that threaten the intellectual and creative property of writers and performers.
But why should such regulations be confined to Hollywood? Shouldn’t workers of other types be equally concerned, and equally protected?
The Dead Professors Society
At Concordia University in Canada in early 2021, Aaron Ansuini was enjoying his remotely delivered module on the history of Canadian art. His teacher was excellent, one of the most respected and well-known art historians in the country. Even on pre-recorded lectures, his passion and personality shone through.
But sometime during the term, Aaron asked his lecturer a question via email, and received no reply. After a bit of Googling, he discovered why: Professor Francois-Marc Gagnon’s funeral had occurred two years prior. The students in the remotely delivered module, it seems, had not been officially informed.
Amid widespread budgetary and staffing constraints in academia, the use of deceased faculty members’ lecture materials represents a solution, and there’s no shortage of raw materials. During the Covid-19 lockdowns, many universities insisted on recording remotely delivered lectures, which then became the university property, to do with as they wished.
Taking this a step further, and use a famous academic’s body of work, correspondence, notes, teaching materials, and recorded talks to train an AI, and less seasoned lecturers could face real competition. From whom would you rather learn physics: a realistic, interactive simulation of Stephen Hawking, or a greenhorn who’s just received their doctorate?
And the halls of academia aren’t the only corridors in which digital ghosts could continue to lurk: the C-suite offices could be next. Hossein Rahnama is just one entrepreneur working with control-hungry CEOs desirous of hanging onto the reins of power even after they shuffle off this mortal coil. In 2018, Rahnama described to MIT Technology Review how he was building avatars of chief executives to allow them to continue posthumously as virtual consultants. More recently, it was reported that while it wasn’t clear if Elon Musk, CEO of the company formerly known as Twitter, had signed up, he’s shown interest in the concept of living on forever in virtual form.
Of course, within the algorithms animating any famous dead professor or influential CEO would be encoded not just the genius and the good but the bias and the bad: outmoded ideas, mistakes, unsavoury prejudices, hubris, and various ‘isms.’ Many theories and business practices of yore should be moved past, but nevertheless some charismatic thought-leaders remain dominant in the cultural imagination long after they should have passed their sell-by dates and ceded power and opportunity to younger generations.
In The Life of Reason (1905), George Santayana said: ‘Those who cannot remember the past are doomed to repeat it.’ But when the past never leaves, are we doomed to never escape it?
AI advances rapidly, regulation lags behind it, and we continue to die leaving behind increasingly vast swathes of rich digital remains, ripe for impersonation, replication, and commercial exploitation. Not long ago, the questions of what happens to people’s data when they die, and who has the right and power to limit its uses, were seen as quirky, even morbid academic concerns. They aren’t so niche anymore.
Further Reading
Kasket, E. (2020). All the Ghosts in the Machine: The Digital Afterlife of Your Personal Data. London: Robinson/Little Brown.
Öhman, C. (2024). The Afterlife of Data: What Happens to Your Information When You Die and Why You Should Care. .Chicago: The University of Chicago Press.
Stokes, P. (2021). Digital Souls: A Philosophy of Online Death. London: Bloomsbury.