The Question

A holographic figure representing a digital afterlife persona

In 2017, journalist James Vlahos watched his father John die of lung cancer. Before he died, James recorded hours of conversations with him — stories, memories, opinions, jokes — then built "Dadbot", a chatbot that answered questions using his father's own words. He wrote about it in Wired. It wasn't his father. But it wasn't nothing either. Vlahos went on to found HereAfter AI.

What was an experiment in 2017 is a small product category in 2026. HereAfter AI lets people record life stories so relatives can later ask questions and hear recorded answers. StoryFile built interactive video interviews, including with William Shatner and with Holocaust survivors, but filed for bankruptcy protection in 2024 with $10.5 million in liabilities. In South Korea, a 2020 MBC television documentary, Meeting You, used virtual reality to recreate a mother's seven-year-old daughter who had died; footage of the mother reaching out to the girl was watched by millions online.

What the Evidence Shows

The digital afterlife industry — sometimes called "grief tech" — spans several kinds of product. Voice cloning services can reproduce a person's voice from short samples. Chatbots trained on someone's emails, texts, and posts can sustain plausible conversations. Video avatars lip-sync to generated speech. The idea runs deep in the AI companion business: Replika began as a chatbot its founder, Eugenia Kuyda, built from the messages of a friend who had died.

Eternos, a US start-up, pitches a service that lets people build an AI version of themselves while alive, trained through hours of guided conversation, so that family members can talk to it after they die. Project December, created by game designer Jason Rohrer, drew attention in 2021 after a man used it to simulate conversations with his late fiancée. In China, companies offer cheap avatars of dead relatives built from photos and voice clips.

"People might develop strong emotional bonds with such simulations, which will make them particularly vulnerable to manipulation."

— Dr Tomasz Hollanek, Leverhulme Centre for the Future of Intelligence, University of Cambridge, May 2024

The underlying technology is improving fast. Large language models trained on a person's writing can imitate vocabulary, opinions, and humour, and combined with voice and video synthesis the result can be startlingly convincing. That is what worries researchers. In 2024, ethicists at the University of Cambridge's Leverhulme Centre for the Future of Intelligence called the field "high risk" and warned that "deadbots" could be used to advertise to grieving relatives or could distress people who never agreed to be simulated.

The question is no longer whether we can build digital replicas of the dead. The question is who controls them — and whether anyone asked the person first.

Why This Is Happening

We are the most documented generation in history. People now leave behind huge volumes of digital material — emails, texts, social posts, voice messages, and video. Someone who died recently likely left enough to train a convincing imitation of their writing style, without ever intending to.

Grief is universal. Roughly 60 million people die worldwide each year, and each leaves people who mourn them. Digital afterlife companies sit at the intersection of that need and the AI boom. But demand has been hard to turn into a business: several start-ups in the field have struggled to find paying customers.

Cultural attitudes toward death are changing. Younger generations, raised with social media, are used to a persistent digital self, and memorial pages on social networks are now common. Whether that comfort extends to talking with a simulation of the dead is untested at scale.


What Could Happen

Digital afterlife becomes a mainstream consumer service within a decade Possible · 25%

Driven by falling AI costs and growing cultural acceptance, digital afterlife services achieve widespread adoption similar to online memorial pages today. Major technology platforms add legacy features to their products, laws establish consent and data-ownership rules, and grief therapists develop protocols for using digital personas.

The industry stays a niche, shaped by consent rules Most likely · 50%

Laws on consent and digital replicas spread. California's AB 1836, in force since January 2025, already requires an estate's consent for digital replicas of dead performers in films and recordings. Similar rules, plus transparency requirements under the EU's AI Act, keep the industry legitimate but small, used mainly by people who plan their own digital legacy.

The technology fragments grief in deeply harmful ways Possible · 25%

Psychologists document a new pathology: digital grief dependency, in which bereaved individuals become unable to process the finality of death because they are in continuous simulated contact with the deceased. A significant minority of users develop what clinicians describe as "complicated grief amplification" — unable to move forward emotionally precisely because the digital persona feels too real. Public health responses treat digital afterlife services as a potential mental health risk, similar to how social media's effect on adolescent mental health is now viewed.

Our Assessment
We assign 25% probability — unlikely that digital afterlife services become a recognized and widely used category of consumer technology within the next decade. We have lowered this forecast from 76%. Our earlier version cited an MIT study, a YouGov survey, investment and sign-up figures, and a quote we could not verify, and it misdescribed several companies. The technology is real and getting better, and some families will find comfort in it. But leading firms have struggled financially, researchers warn of manipulation and grief harms, and consent laws are tightening. A meaningful niche is far likelier than a mainstream habit.

What Can We Do

A person reviewing digital legacy settings on a tablet

Whether you find the prospect of digital afterlife comforting or deeply unsettling, it is arriving whether you plan for it or not. Individuals, families, legislators, and technologists all have decisions to make — and making them proactively beats having them made by default.

Document your wishes explicitly. Just as you might include instructions about organ donation or funeral arrangements in your will, consider stating whether you want to be digitally recreated after death. Many major platforms let you name a legacy contact or choose what happens to your account. If you do not want to be replicated, say so in writing now.

Archive intentionally, not accidentally. If you do want to leave a digital legacy, recorded stories and letters made on purpose capture far more of who you are than years of incidental social media posts — and they are valuable to your family whether or not anyone ever builds an AI from them.

Approach digital personas in grief with clinical support. If you are considering using a digital afterlife service to interact with a deceased loved one, do so with the involvement of a grief counselor or therapist who can help contextualize the experience. These tools are not substitutes for grief — they are supplements to it, and the distinction matters enormously for psychological outcomes.

Engage with the regulatory conversation. The rules on who owns your digital identity after death, who can create replicas without consent, and what protections bereaved families have are being written now. Groups such as the Electronic Frontier Foundation are active in these debates, and public input will shape whether the technology develops with meaningful safeguards.