Digital Immortality: When AI Lets Us Talk to the Dead
What if you could talk to a loved one who has passed away? Hear their voice, recognize their mannerisms, and pick up a conversation cut short by death? Behind the promise of reconnecting with those we've lost through digital immortality, lie ethical, psychological, and legal questions we have yet to fully answer.
Digital Immortality: From Science Fiction to Reality
Throughout our lives, we leave behind a digital trail of photographs, videos, text messages, emails, social media posts, and voice recordings. These fragments of our lives remain long after we're gone.
Thanks to advances in generative artificial intelligence, this data can now be used to create a digital representation of someone, capable of imitating their voice, appearance, and communication patterns.
These representations are sometimes called griefbots or deadbots. They work in different ways. Some technologies generate videos in which deceased individuals appear to speak to their loved ones. Others allow people to have conversations with AI-powered simulations of the deceased, using chatbots trained on their personal data.
One fundamental distinction must be made, however. These technologies don't bring anyone back to life. They create simulations based on available information.
The responses they generate don't necessarily reflect what the person would have actually said or thought. This distinction becomes particularly important when we consider how these interactions might affect people who are grieving.
Companies Are Already Recreating Our Loved Ones
This market has moved well beyond the conceptual stage. In South Korea, a company called Vaice creates videos of deceased individuals using photographs and voice samples. According to an Associated Press report, the company says it serves approximately 300 customers a month. A three- to five-minute video costs about US$390.
The report tells the story of a man who commissioned a video of his late grandfather as a gift for his father. In the recreation, the grandfather speaks lovingly and apologizes for certain events from the past. His son is overcome with emotion and bursts into tears.
An important detail worth mentioning is that the message was written by the grandson. AI hadn't somehow recovered the deceased man's actual thoughts. It had simply given those words his voice and face.
This distinction is essential to understanding what these technologies can truly offer. And their possibilities extend well beyond prerecorded videos. Companies are also exploring interactive avatars that can answer questions, recount memories, and hold conversations with a deceased person's loved ones.
Even tech giants are interested in such possibilities. In December 2025, Meta was granted a patent describing a system capable of simulating a user's activity on social media, including after their death, using artificial intelligence. The technology could potentially generate posts and interactions that mimic those of the deceased user. Meta has since stated that it has no plans to commercialize the technology.
How Does AI Imitating a Deceased Loved One Affect the Grieving Process?
For some people, hearing a familiar voice or recognizing someone's way of speaking can bring comfort. For others, the experience could reopen emotional wounds or make it harder to come to terms with their loss. Researchers are beginning to explore these reactions.
A study reported by Science News examined how 16 volunteers interacted with conversational simulations of deceased loved ones. Participants generally described the experience positively, even when the AI invented details about the person who had died.
Interestingly, factual inaccuracies seemed to bother participants less than mistakes in tone or manner of speaking. In other words, a simulation can create a convincing sense of someone's presence without accurately representing who they were. This phenomenon echoes something we explored in Love and AI: Can You Fall in Love With a Chatbot?:
"Our brains don't stop to verify the true nature of the person—or entity—we're interacting with. They respond to signals of connection, and those signals are convincing."
In the context of grief, this distinction becomes particularly unsettling. A familiar voice, recognizable expressions, and affectionate responses can trigger a genuine emotional reaction, even when the person we believe we're reconnecting with is nothing more than a digital simulation.
These findings should nevertheless be interpreted with caution. The participants volunteered for the study and were open to the experience. Furthermore, a researcher supervised the chatbots' responses. The results therefore don't establish that digital ghosts help people cope with grief in general.
An article published in the Journal of Medical Internet Research also points out that research into the effectiveness and risks of these technologies remains limited.
Some experts see potential for their use alongside therapeutic or memorial practices. Others worry that prolonged interactions with realistic simulations could encourage emotional dependency or blur the line between a person's actual absence and their artificial presence.
The question, then, isn't simply whether the technology works. It's how it changes the way we experience the loss of someone we love.
The Risk of Losing Someone a Second Time
Imagine someone who has spent years talking every day to a digital replica of their deceased spouse. Then, one day, the company hosting the service shuts down.
The avatar disappears.
That person could feel as though they've lost their loved one all over again. This scenario is among the concerns raised in an October 2026 investigation published by Vox, which examines research by Katarzyna Nowaczyk-Basińska and Tomasz Hollanek of the University of Cambridge.
The researchers are interested not only in the possibility of a "second loss," but also in the control companies could exert over these artificial relationships. Unlike our personal memories, which belong to us, digital avatars often depend on commercial infrastructure.
Access to a virtual loved one could therefore depend on an ongoing subscription, a platform's terms of service, or the financial survival of the company providing the technology.
What happens if that company decides to change the avatar's behaviour? Insert advertisements into its conversations? Or use the information collected for other purposes?
These scenarios are explored in a Cambridge research paper on the ethical implications of posthumous avatars. The issue goes far beyond technology. It raises questions about the commercialization of grief and the vulnerability of people going through an emotionally difficult time.
Who Has the Right to Digitally Recreate Someone Who Has Died?
Another fundamental issue is consent. Consider a woman who dies without ever expressing a wish to be represented by artificial intelligence. A few years later, one of her children decides to create an avatar using her messages, photographs, and voice recordings. His brother finds the idea comforting. His sister, on the other hand, believes it disrespects their mother's memory.
Who gets the final say?
What happens if the conversations used to create the avatar contain private information about other people?
Researchers are examining the rights of deceased individuals, consent from family members, data retention, and the limits of commercial use. A study published in the Journal of Responsible Technology proposes two guiding principles for regulating griefbots: consent and the prevention of emotional dependency.
The authors recommend establishing safeguards around the creation, use, and deletion of avatars, clearly identifying simulations as such, and preventing features that could encourage excessive emotional attachment.
Another research paper published in June 2026 proposes a framework covering nine dimensions, including consent, data sources, the accuracy of the simulation, its intended purpose, and system governance. These studies show that ethical concerns cannot be resolved by simply checking an "I agree" box in a service's terms and conditions.
What About Digital Immortality in Quebec?
The issue is particularly relevant when it comes to protecting personal information.
Quebec's Commission d'accès à l'information explains that, under certain conditions, individuals may obtain personal information about a deceased loved one, particularly when that information could help them through the grieving process.
However, the right to access such information does not constitute blanket permission to reproduce someone's personality, voice, or likeness using artificial intelligence. Questions surrounding personality rights, privacy, and consent after death remain complex and require legal analysis based on the circumstances of each case.
Quebec Researchers Are Exploring Digital Immortality
Quebec is also contributing to the discussion surrounding these emerging technologies.
The International Observatory on the Societal Impacts of AI and Digital Technology (OBVIA) is conducting a research project entitled Artificial Immortality: Ethical, Legal, and Artistic Perspectives.
Led by Emmanuelle Marceau, a philosophy professor at Cégep du Vieux Montréal, the project brings together expertise in philosophy, law, and artistic creation. An article published by the Fonds de recherche du Québec in 2025 highlighted the many questions raised by the use of digital avatars representing deceased individuals. The research has also led to Artificial Immortality Is on Our Doorstep: Cure-All or Danger?, a reflection on the ethical risks associated with these emerging practices. Marceau's research examines the potential consequences for the grieving process, individual rights, and our collective relationship with death. This interdisciplinary approach matters because digital immortality is not simply a matter of computer science. It also touches on psychology, law, philosophy, culture, and societal values.
Should We Be Wary of Digital Ghosts?
It would be far too simplistic to portray these technologies as either an extraordinary breakthrough or an inherently dangerous development. Some people may find comfort in a one-time experience, such as watching a memorial video or listening to a message recorded by a loved one during their lifetime. Others may develop an ongoing relationship with a conversational avatar whose responses become increasingly difficult to distinguish from those of a real person.
These two experiences don't carry the same risks.
Before using a digital immortality service, several questions are worth asking:
Did the person consent to the use of their image, voice, and personal data?
Do users clearly understand that they're interacting with a simulation capable of making things up?
How is personal information stored, used, and eventually deleted?
Can users discontinue the service without permanently losing the data or memories they've entrusted to the company?
Are safeguards in place to protect against commercial manipulation and emotional dependency?
These precautions become especially important when children or vulnerable individuals are involved. The goal should be to preserve the dignity of those who have died and protect the well-being of their loved ones, rather than artificially extending interactions at any cost.
AI Can Imitate a Person, But Can It Preserve What Made Them Unique?
Digital immortality ultimately confronts us with a question that goes beyond the capabilities of artificial intelligence.
What truly defines a person?
Their voice? Their memories? The way they speak? Their opinions? Their ability to surprise us?
AI can reproduce some of these characteristics using data. It can even generate entirely new responses that create the impression of a natural conversation. But that doesn't mean it possesses the deceased person's lived experiences, consciousness, or subjective awareness.
A convincing resemblance doesn't mean the person somehow lives on. Yet this doesn't diminish the emotional value that some people may find in these simulations. After all, our relationships with the memories, belongings, and images of those we love are deeply personal.
Unlike a photograph or an old voice message, an AI-powered avatar can generate entirely new statements. It can attribute words to someone who never actually said them. Perhaps that's where the real transformation lies. For the first time, our digital memories can talk back to us. This possibility forces us to consider not only what technology makes possible, but also what we truly want to do with it. The question is no longer simply whether we'll be able to keep talking to our loved ones after they die. It's also about deciding who should have the right to speak on their behalf.
Understanding AI to Better Navigate Its Implications
Artificial intelligence is transforming the way we work, communicate, and even maintain relationships. Keeping up with these changes means understanding what these technologies can actually do, where their limitations lie, and the questions they raise.
Info IA Québec offers conferences on artificial intelligence to help the general public, professionals, and businesses develop a well-informed understanding of AI and its applications.
Natasha Tatta, C. Tr., trad. a., réd. a. I'm a certified translator, bilingual content writer, infopreneur, and GenAI consultant. I work at the intersection of language, content marketing, and technology, helping professionals and organizations make practical use of AI and communicate with clarity, purpose, and impact.
