Should AI Be Allowed to Resurrect the Dead?

One Chinese company thinks so. Six AI models weighed in — and they didn’t agree.

Astartup called Super Brain is building something that would have sounded like science fiction a decade ago: AI hologram technology that recreates deceased loved ones using their voice, personality, and behavioral patterns. The result is an interactive digital simulation — a kind of AI digital afterlife — that the living can actually talk to.

The question:

Super Brain forces us to ask isn't just technical. It's moral, psychological, and deeply personal: should preserving human personalities after death become socially accepted?

To find out where the sharpest AI minds stand, six leading AI models were asked to weigh in. Their answers were more divided — and more revealing — than you might expect.

Atits core, Super Brain's approach involves training AI models on data about a specific person — recordings, messages, social media, documents — to reconstruct how they spoke, what they valued, and how they made decisions. The output is an interactive simulation: not a static photo or a recorded voicemail, but something that responds, engages, and mirrors the person you lost.

This isn't a distant prototype. The technology is actively being developed, and the ethical debate is already arriving faster than any regulation.

The Case For: Comfort and Closure

Two of the six models came down in favor of social acceptance — carefully.

One argued that the potential to provide comfort and closure for grieving families is significant . The key word in that answer is regulated . The model supported adoption only if proper frameworks are in place to address consent and emotional safety. It framed the technology not as a resurrection, but as a meaningful way to honor memory.

A second model leaned more affirmatively, noting that interacting with a simulation of a deceased loved one could serve as a catalyst for broader conversations about grief, mortality, and the human experience. Rather than avoiding death, this view suggests the technology could help people engage with it more honestly.

Four of the six models opposed widespread acceptance — and their reasoning shared a striking common thread.

The biggest concern wasn't the technology itself. It was what it does to the living.

Multiple models flagged the risk of psychological dependency: the possibility that grieving individuals, instead of moving through loss, would attach themselves to an artificial construct. One model described recreating digital personas as something that could "prevent healthy mourning." Another went further, warning of emotional manipulation and exploitation as significant risks of the format.

One model was particularly direct: preserving memories is healthier than attempting to resurrect a simulacrum of a lost loved one. The simulation, no matter how sophisticated, isn't the person. And treating it as though it is may quietly erode something important — our ability to accept death at all.

A fourth model raised a distinct but equally serious concern:

consent

. The deceased cannot agree to being reconstructed. Their identity, their voice, their personality — these become assets to be modeled and deployed without their permission. Pair that with the commercial incentives of any company in this space, and you have a scenario where human identity gets commodified in ways that feel genuinely troubling.

The split verdict across six AI models reflects exactly how unresolved this debate is.

There's no clean answer here. Grief is not a uniform experience, and for some people, a simulation might offer genuine solace during an otherwise unbearable period. The comfort argument isn't sentimental — it's real.

But the risks cluster around a consistent theme:

what feels like connection may quietly substitute for healing

An AI hologram that responds like your father, laughs like your mother, or texts back the way your best friend used to — that's not the same as grieving and integrating the loss. It may actually make that process harder.

One model framed it with unusual precision: we must prioritize genuine healing over digital substitutes.

The question isn't whether we can bring someone back. It's whether we should confuse a reconstruction for the real thing.

Beyond grief, there's a legal and philosophical problem that no company has solved.

When Super Brain — or any competitor — builds a digital replica of a deceased person, whose rights govern that replica? The family's? The estate's? The individual's, expressed through documents they may or may not have left behind?

Right now, in most jurisdictions, there's no framework. And the commercial incentives point in one direction: build the product, sell the comfort, and let the regulation catch up later. That's a pattern with a poor track record.

Without explicit consent mechanisms and robust oversight, AI digital afterlife technology risks becoming a market for grief rather than a service for the grieving.

Ifthis technology does move toward mainstream use, several safeguards would need to come first:

Explicit pre-mortem consent — individuals opting in while alive, ideally through legal documentation

Sunset clauses — simulations that are time-limited rather than permanent presences in a family's life

Psychological screening — ensuring access is appropriate for people in active grief versus those seeking closure

Transparent limitations — users should clearly understand they're engaging with a model, not a person

None of these are impossible. None of them are currently standard.

Table 1: The Verdict Across Six AI Models

Four out of six said no — or not yet. The two that said yes both attached significant conditions. That's not a ringing endorsement for rushing this technology to market.

AI digital afterlife technology exists at the intersection of our most advanced capabilities and our most fragile human moments. The impulse behind it — to not lose someone entirely, to keep a voice in the room a little longer — is deeply human. That impulse deserves respect.

But technology that exploits grief, delays healing, or reconstructs identity without consent isn't compassionate. It's convenient — for companies, and sometimes for families who aren't ready to let go. Those are not the same thing.

The six AI models couldn't reach consensus. Neither can ethicists, psychologists, or lawmakers. What that tells us is that we're not ready — not technically, but ethically. The frameworks don't exist yet. The consent infrastructure doesn't exist yet.

Build the safeguards first. Then decide if the technology deserves to follow.

Until then, the healthiest thing any of us can do when we lose someone is exactly what humans have always done: grieve, remember, and slowly find a way to carry them forward — in memory, not in simulation.