Resisting immortality: do we have the right to remain dead?
By David Falls
Imagine hearing the voice of someone you love, years after their death, answering questions they were never asked. Artificial intelligence makes this increasingly possible. But is it a comforting way to remember someone, or a troubling violation of their autonomy? David Falls explores the ethics of digital resurrection.
David lives in Queen Creek, Arizona, and he is the author of God’s AI Reckoning: The Final Revelation (2025) and The Great Silence: What Remains After Belief (2026).
On the resurrection of Joaquin Oliver
On what would have been his twenty-fifth birthday, Joaquin Oliver gave a video interview. He spoke about gun violence, basketball and some of his favourite films. But Joaquin had been dead for more than seven years. He was seventeen when he was murdered in the 2018 school shooting in Parkland, Florida. The figure answering the questions was an AI recreation made from his voice, image, writings and social media posts.
His parents created the AI version of Joaquin as part of their campaign against gun violence. Their purpose was neither commercial nor frivolous. They wanted their son's views to remain part of a public argument that his murder had given them every reason to pursue. To them, the technology offered another way to preserve Joaquin's memory and continue work they believed reflected his values.
The interview nevertheless produced an uneasy question. The decision had been made by people who loved Joaquin, but it had not been made by Joaquin himself. His parents had access to his photographs, writings and recorded voice. Whether that gave them the right to approve a new version of Joaquin that would answer questions he had never heard was far less clear.
Reactions to the interview were sharply divided. Some people found it moving. Others considered it disturbing or even ghoulish. Neither response is difficult to understand. Grief makes the desire to hear a familiar voice again intensely human. Yet there is something troubling about watching a dead person appear to express new thoughts, however carefully those thoughts have been assembled from what the person once said and believed.
Some troubling ethical questions
This is where the idea of an AI afterlife becomes more than a question about technology. It becomes a question about human autonomy. Who should decide whether a person continues to speak after death? Should a spouse, parent or child be able to make that decision? Does legal ownership of someone's recordings confer the moral right to generate new statements in that person's name? And what happens when the replica belongs not to the family but to the company operating it?
Whatever emotional value the interaction may hold, it does more than preserve what the deceased left behind. It produces new material under someone else's control. If autonomy and dignity matter during life, they should not disappear at death. That may require recognising a right that previous generations had little reason to consider: the right not to be digitally resurrected.
Why would anyone want this?
It's easy to judge such creations from a distance. Speaking with an AI version of the dead may appear morbid to someone who is not grieving. To someone who still reaches for the telephone before remembering that the person will not answer, the attraction may be immediate. Not every digital recreation is equally troubling. Someone might record stories, memories and messages while still alive, knowing that their family will later be able to hear them. An interactive system could organise those recordings and help relatives find a particular story or piece of advice. Used that way, AI would function as a new kind of family archive. The person chose to create it, understood its purpose and determined what it contained.
A system built after death is different. It may be assembled from private messages, social media posts, videos and recordings that were never intended for that purpose. Instead of retrieving answers the person actually gave, it can generate new ones. And the difference may not be obvious to someone using it, particularly when every response arrives in a familiar voice. Even then, the comfort may be genuine. A person might use the replica to recall forgotten stories, introduce a child to a grandparent who died before the child was born, or say things that were left unsaid. The fact that the response comes from a machine does not make the mourner's feelings artificial. Human beings have always formed emotional attachments to photographs, letters, possessions and places associated with the people they have lost.
Nor should we assume that there is one correct way to grieve. Some people keep loved one's voicemails for years. Others find recordings too painful and remove them. An AI recreation might help one person remember while preventing another from accepting the loss. The technology cannot determine which response is healthy.
“The people who benefit from the replica are not the people being replicated.”
A humanist approach should take these emotional needs seriously. Dismissing every digital replica as grotesque would ignore the love and pain that often motivate its creation. But compassion for the living cannot settle the entire question. Something may help a grieving family while using the identity of the deceased in a way that person would never have accepted. That's the difficulty at the centre of the AI afterlife. The people who benefit from the replica are not the people being replicated.
Consent of the dead
We already accept that a person's wishes can extend beyond death. People decide what should happen to their property, their bodies and their remains. They may choose whether to donate organs, where to be buried, and who may gain access to private records. It would be strange to respect those decisions while assuming that a person's identity becomes available for reconstruction as soon as life ends. The controversy surrounding the documentary Roadrunner showed how quickly the issue can arise. The film used artificial intelligence to recreate Anthony Bourdain's voice and have him speak several sentences that he had written but never recorded. The director said he had received the necessary approval, but Bourdain's widow, Ottavia Busia-Bourdain, disputed the suggestion that she had agreed Bourdain would have accepted it. The words belonged to Bourdain, but the decision to make him speak them after his death did not.
That case involved words Bourdain had actually written. A conversational replica goes much further. It may be asked about political events, family disputes or social changes that occurred years after the person died. Its answers will be based partly on evidence from the person's life and partly on decisions made by developers about how the system should respond. Yet every answer may still appear beneath the dead person's name. Meaningful consent would therefore have to be specific. Agreeing to preserve photographs or recordings is not the same as agreeing to an interactive recreation. Allowing a company to store social media posts is not permission to use them to build a personality. Even someone who approves a private family archive may not want a replica appearing publicly, giving interviews or endorsing causes and products.
Consent must also define the limits. A person ought to be able to decide who may interact with the replica, what information it can use, whether it may generate new opinions and how long it may continue to operate. Permission to create one version should not become permanent permission for future companies to produce others. The most defensible rule is straightforward: no interactive recreation without clear permission from the person being recreated. Family approval may be important, but it should not replace the choice of the individual whose voice, face and personality are being used. Love does not grant ownership of another person's identity.
This rule would also prevent a loving family from creating a replica when someone died without leaving instructions, as Joaquin Oliver did. That is a painful consequence, but the absence of instructions cannot be treated as permission. Without that protection, the moment a person can no longer object may become the moment when others gain the greatest freedom to speak in that person's name.
Who controls the replica?
Imagine that a woman speaks regularly with an AI version of her dead mother. It recalls family stories, asks about the grandchildren and answers in a voice the woman has known all her life. Yet neither mother nor daughter controls it. A technology company stores it, operates it and determines how it responds. The family may own the photographs, messages and recordings used to create the replica, while possessing little control over the model created from them. The company may update its software, change its safety rules or discontinue the service. A future version might respond differently from the one the family first encountered. If the business is sold, the information used to recreate the deceased may pass to a new owner. If the subscription becomes too expensive, continued access could depend on the family's ability to pay.
The business model matters. A company that earns money from continued interaction has an incentive to keep people returning. It may encourage longer conversations, send reminders that the replica is waiting or charge for additional access to its voice and memories. Grief then becomes a source of revenue, and the digital presence of someone we loved becomes a service that can be upgraded, restricted or withdrawn.
Control can also divide families. One relative may find it comforting while another believes its existence violates the deceased. A spouse may want to preserve it, while an adult child wants it deleted. They may disagree about whether it should discuss private matters, communicate with children or appear publicly. The person whose identity is being disputed can no longer resolve any of these questions.
Public figures present a greater risk. A posthumous replica could promote a film, support a political cause or endorse a product released decades after the person died. Even when its answers are based on earlier views, no one can know whether the individual would have changed their mind. The replica may preserve a recognisable personality while placing it in the service of people and institutions the deceased never chose.
Responsible use therefore requires more than initial permission. A person should be able to designate who controls the replica, restrict how it may be used and prohibit commercial or political endorsements. Families should know how responses are generated, what information the system uses and whether the company retains it. There must also be a clear way to delete the replica permanently.
Without such protections, an AI afterlife would not belong to the person it appears to preserve. It would belong to whoever controls the software.
“No one should be digitally resurrected unless they chose that future for themselves.”
A replica Is not an afterlife
Even with consent and responsible control, one distinction must remain clear. A digital replica is not evidence that the original person has survived. It may reproduce a voice, imitate a manner of speaking and provide responses consistent with years of personal information. None of that demonstrates that the person's awareness or experience continues within it.
This conclusion does not depend on belief in a soul. A humanist can recognise the difference between preserving information about a life and preserving the individual who lived it. The replica has no memory of childhood in the sense that the person did. It has information about that childhood. It may describe a marriage, a friendship or a moment of fear, but it did not experience any of them. The ability to speak as someone is not the same as being that person.
People using these systems should be reminded that the replies are generated, particularly when they offer advice or comments on events that occurred after the death. Companies should not advertise these systems as a form of immortality or encourage users to believe that consciousness has been transferred. Whatever value the interaction may hold, the deceased has not returned.
Used carefully, an AI recreation could be a meaningful form of remembrance. It might preserve family history, share stories with later generations or allow someone to leave messages for important moments in the lives of those they love. But it should exist only when the person knowingly chose it, understood what it could become and retained authority over how it would be used.
The right to create such a version of ourselves should be accompanied by an equally important right to refuse. We should be free to leave recordings, memories or even an interactive likeness for the people who survive us. We should also be free to let our voices fall silent.
Conclusion
A humane society should not treat death as permission to take control of an identity. If technology gives us the power to make the dead appear to speak, restraint matters as much as invention. No one should be digitally resurrected unless they chose that future for themselves.
Comments