Seance in AI age: the ethics of 'grief tech'
In the age of artificial intelligence (AI), death no longer has to be the end of a conversation. A recent New York Times Magazine article explored the rise of "grief technology" — AI avatars designed to digitally resurrect the voices, personalities and memories of the deceased. These hyperrealistic “ghosts in the machine” offer comfort, connection and, in some cases, closure. But they also raise profound ethical concerns. Are we enhancing the human experience of mourning — or disrupting a sacred process? Is this technology a tool for healing, or a dangerous manipulation of our most vulnerable moments?
The latter case seemed especially relevant when I watched a Korean TV show that brought back AI-generated virtual reality holograms of dead loved ones. It’s difficult to not be affected by a mom crying when seeing her lost little girl again, filled with despairing grief over the loss and a desperate need to believe. However, it was distasteful as much as it was compelling.
Grief tech is an expanding field. Companies now offer services that allow individuals to upload voice memos, chat logs, photos and video recordings to create AI models that simulate loved ones. These avatars can answer questions, share memories and even hold seemingly spontaneous conversations. What once required a seance now takes an app.
On an individual level, the benefits can feel deeply personal and profound. For those who lose someone suddenly, an AI avatar can serve as a bridge — helping them process trauma, say goodbye or even revisit treasured moments. But beneath this emotional promise lies a tangle of psychological risks. Grief, by its nature, is meant to be finite. It is a painful but necessary journey toward acceptance. AI avatars can complicate that process, prolonging denial and turning grief into a loop. Instead of learning to live with loss, users may retreat into digital re-creations, clinging to simulated conversations that may feel real, but are ultimately hollow and one-sided. They are talking to a pattern, not a person.
There’s also the question of authenticity. No AI — no matter how sophisticated — can truly replicate the essence of a human being. These avatars may speak in familiar tones and recall private jokes, but their responses are generated, not felt. A relationship is not reductive. That is, it’s not just two wholly formed individual agencies coming together to interact for a while. It’s two agencies that come together to meld into a new one that remake and redefine who they are to each other. It’s a dynamic, living process, one which AI cannot replicate. In trying to preserve the dead, we may end up distorting ourselves.
The ethical dilemmas deepen when we consider grief tech on a societal scale. If this technology becomes normalized, how might it reshape our understanding of death and memory? Funerals and photo albums might be replaced by digital companions. Will we become less inclined to let go of our losses, always keeping a foot in the past? And what happens to the boundaries between the living and the dead?
There are even thornier concerns around consent and privacy. Many of these avatars are created using data the deceased never agreed to share in this way. Who owns a person’s voice after they’re gone? Should families be allowed to bring someone back digitally without that person’s explicit permission? These questions demand urgent legal and ethical frameworks that do not yet exist.
Moreover, grief tech opens the door to exploitation. What begins as comfort could easily be monetized, gamified or manipulated. Imagine targeted ads delivered through your late grandmother’s voice, or political messages voiced by a deceased celebrity. The commercialization of the dead is not just disrespectful — it’s dangerous.
Of course, every new technology drives changes that we have to adjust to. But with something as transformative as AI, we must proceed with caution. More than anything else, death is what makes us human. At its core, grief tech touches on that very humanity: to love, to lose, to remember and to forget. We cannot allow algorithms to define those sacred experiences without careful deliberation and consensus.
We need thoughtful dialogue between technologists, ethicists, mental health professionals and the public. We need clear standards for data use, informed consent and emotional harm mitigation. We need to teach digital literacy that includes not just how to use grief tech, but when to put it down.
We should not fear the evolution of mourning. Humanity has always created rituals to stay connected to the dead — tombs, portraits, letters, recordings. AI avatars are merely the newest form. But unlike past memorials, these ones talk back. That is a difference worth scrutinizing.
As we stand at the intersection of technology and mortality, we must ask not only what we can do, but what we should do. The dead deserve dignity. The living deserve clarity. And grief, no matter how painful, deserves its rightful closure.
Jason Lim (jasonlim@msn.com) is a Washington-based expert on innovation, leadership and organizational culture. The views expressed in this article are the author’s own and do not reflect The Korea Times’ editorial stance.