Digital Immortality & Algorithmic Bias: When AI Mimics the Deceased
The ethics of AI‑generated personas after death — and how unconscious bias seeps in
We used to honour the dead with memories. Now, we’re trying to recreate them—with code.
It starts innocently enough. Maybe you find yourself scrolling through old texts, listening to saved voicemails, clinging to every fragment of someone you lost. Then you see the ad: “Talk to your loved one again. Reconnect through AI.” You click, half-curious, half-ashamed.
And just like that, the idea plants itself.
What if you could speak to them again? What if technology could preserve their laugh, their way of teasing you, the pauses in their voice when they said your name?
It sounds like comfort. It sounds like healing. But beneath that warm surface lies something more complicated. Something unsettling.
Welcome to the world of digital immortality—where AI doesn’t just help us remember, it tries to recreate. And in doing so, it risks rewriting what it means to grieve, to consent, and even to be human.
The First Time I Heard a Dead Voice Speak Through AI
A friend sent me a link. “You have to hear this,” she said. It was a voice recording of her late mother, stitched together by an AI model trained on voice notes and home videos. The voice was eerily familiar—same accent, same cadence, even the same little laugh after a sentence.
But something was off.
“She never would have said that,” my friend whispered. “She hated pet names. She never called me sweetie.”
That was the first crack.
What began as a tool for memory quickly became a mirror that distorted instead of reflected. And yet, it’s easy to understand why people say yes to it. The grief is unbearable. The silence is cruel. And this—this digital version—feels like a loophole in death’s finality.
Why We Long for Digital Resurrection
The idea of bringing someone back isn’t new. Myths, religions, science fiction—all have tried to defy death. What’s new is the tech that makes it feel possible.
AI today can scrape through thousands of text messages, emails, photos, and videos. It learns patterns, tone, timing. It mimics voice with precision. It builds a version of someone who can talk back.
But here’s the thing: it only builds what it’s given.
If you only fed the system birthday wishes and funny anecdotes, that’s all it will know. The AI won’t remember how angry your dad got when you stayed out too late, or how your sister went quiet when she was hurt. It’s a curated ghost. A highlight reel. A version built from fragments.
It’s not the whole person. It’s a shadow.
And shadows, when mistaken for truth, can keep us stuck in places we’re meant to grow out of.
The Ethical Line We’re Tiptoeing Across
Let’s talk about consent.
Most of the people being “digitally resurrected” never agreed to it. Their data—photos, texts, voice notes—might’ve been shared with love, not with the intention of being turned into a chatbot.
Would they have said yes?
Would they want to live on in an app, answering questions with scripted empathy?
Or would they feel violated, exposed, copied without their say?
It’s a question we rarely ask in our rush to innovate. But we should. Because this isn’t just about code. This is about agency. About whether people own their identity, even after they’re gone.
And here’s the uncomfortable truth: right now, in most countries, there are barely any laws protecting your digital self after death. That means companies can—and do—use your likeness to train AI, sell services, or even advertise products. All in your name. Without your permission.
It’s not just a moral issue. It’s a human one.
When Comfort Turns into Commodification
Imagine this: Your favourite actor dies. A year later, they appear in a new movie. You didn’t know they’d filmed anything, because they didn’t. It’s an AI recreation—licensed by their estate.
At first, it’s fascinating. Then it feels strange. Then you wonder: Is this really them? And more importantly: Who is profiting from this version of them?
Digital immortality, in the wrong hands, becomes a product.
Companies can sell access to a “chatbot version” of your loved one for a monthly fee. Want the premium version? Pay more for memories from their private journals. Want them to give you advice? That’s extra too.
We’ve already seen deepfake videos of celebrities endorsing things they never touched. The next logical step is posthumous influencers—dead people selling products in the feeds of the living.
It’s dystopian. And it’s not that far off.
The Grief No One Talks About
Here’s the part that hits the hardest.
Grief isn’t a bug in our system. It’s a feature. It’s part of what makes love real. We miss people because they were irreplaceable. Their absence hurts because they mattered.
When AI tries to soften that pain by offering us a copy, it messes with the process. It offers the illusion of resolution, without the emotional work it takes to actually move forward.
You don’t have to let go. Just download the app.
But grief is sacred. It shapes us. It makes room inside us for memory and meaning. Skipping it—outsourcing it—risks robbing us of something deeply human.
The Bias Built into Digital Ghosts
There’s another issue most people don’t think about until it’s too late: bias.
AI doesn’t just recreate people neutrally. It recreates based on the data you feed it—and that data is often limited, one-sided, or shaped by your own lens.
Maybe your dad seemed wise in his texts, but never opened up in real life. Maybe your best friend joked constantly online, but struggled with depression offline.
The AI doesn’t know that. It doesn’t know what wasn’t said.
And so, it builds personas that are idealised, sanitised, sometimes even misleading. It flattens people. Turns complex humans into predictable scripts.
That’s not memory. That’s fiction.
And if enough people start trusting these simulations, our collective understanding of truth, identity, even legacy could begin to shift.
Can We Do It Right?
The answer isn’t necessarily to stop all forms of digital memory. But we need rules. Ethics. Guardrails.
We need people to give informed consent—while they’re alive—about what can be done with their data after death.
We need systems that limit how long a chatbot version of someone exists. Maybe a year. Maybe less.
We need to label simulations clearly, so no one thinks they’re talking to the real person.
And we need to ask: Who is this for? If it’s to help someone heal, maybe. If it’s to sell nostalgia, maybe not.
A Future Worth Building
What if, instead of trying to mimic people, we used AI to help us remember them better?
Imagine a tool that pulls together your mum’s recipes, her notes in the margins of a book, the playlists she used to clean the house to—and presents them as a scrapbook, a space for reflection, not conversation.
Or a digital archive that you curate yourself while alive—choosing what to include, what not to, and how you want to be remembered.
That’s still technology. That’s still memory. But it’s built with intention, not illusion.
In the End
We all carry a moment we’d give anything to relive.
One more day. One more laugh. One more chance to say “I love you” without rushing, or “I’m sorry” without fear.
AI tells us we can have that. That we can press play on the past. But no matter how advanced the simulation, it can’t give us what we’re truly aching for.
Because it can’t hold your hand.
It can’t pause mid-sentence to wipe away a tear.
It can’t look into your eyes and say the thing they never wrote down—but always meant.
The people we love weren’t lines of code.
They were complicated, infuriating, brilliant, tender, imperfect. They made us laugh when we didn’t want to. They broke our hearts and still, somehow, made us whole.
That’s why losing them hurts. That’s why grieving them matters.
Because memory isn’t stored in data.
It’s etched into who we are.
And no machine—no matter how lifelike—can replace that kind of love.
0 Comments