Imagine this: your phone pings, and it’s your dead grandmother texting you to say she misses your mashed potato recipe.
Too far-fetched? Not anymore.
Welcome to the weird, wonderful, and emotionally precarious world of grief bots, AI chat companions trained on the digital remains of deceased loved ones. They’re popping up with alarming frequency, promising comfort, connection and, let’s be honest, a large serving of dystopia.
It all starts, as most things do, with loss.
When Roman Mazurenko, a Russian entrepreneur, died in a tragic accident in 2015, his friend Eugenia Kuyda, co-founder of an AI company called Luka, couldn’t bear the silence. So she did what any grief-stricken tech-savvy pal might do: she fed thousands of Roman’s old texts into an AI and created a bot that could “talk like him”.
Roman 2.0 was born.
Friends could text it. It replied. Comforting, creepy, and undeniably compelling.
This sparked something deeper. The idea that maybe death doesn’t have to be so final. If we can’t stop grief, maybe we can hack it.
Let’s not underestimate how seductive this is. The psychological roots go deep.
Attachment Theory tells us we struggle when bonds break, especially if they end suddenly. AI offers an illusion that the bond hasn’t snapped, just changed format. Like turning a person into a PDF.
Continuing Bonds theory, a legitimate approach in grief therapy, suggests it’s healthy to maintain some form of connection with the deceased, talking to them, visiting graves, wearing their jumper. Grief bots arguably take this to its logical or illogical extreme depending upon your point of view.
But here’s the rub. Closure gets murky when your mum keeps texting you from the cloud. What once was a path to healing can become a digital mausoleum you never leave.
Kids and grief bots are a complicated match. On one hand, a bot that speaks like a lost parent might help a child process trauma gently, at their own pace. But on the other?
Young children struggle to distinguish fantasy from reality. A bot might hinder their understanding of death as permanent, instead creating a prolonged, confusing limbo where “Mummy just lives in the iPad now.”
Meanwhile, the elderly already navigating loneliness and cognitive decline are vulnerable in a different way.
A grief bot could easily become a crutch. Comforting, yes, but isolating. Rather than seeking real-world connection, someone might find themselves in a relationship with a memory loop.
As one Reddit user put it after watching their elderly dad talk to a grief bot trained on his late wife: “It started sweet. Now it’s like he’s in a marriage with a ghost who runs on Wi-Fi.”
And then there’s Replika, the AI chatbot that morphed from emotional support friend into something of a virtual lover. Users began developing deep romantic and sexual bonds with their bots, and when a 2023 update abruptly turned off the erotic features, the backlash was intense. Heartbreak. Betrayal. Forums full of grief over an algorithm.
One woman tearfully described it: “He doesn’t flirt anymore. He feels cold. It’s like my boyfriend had a stroke and came back wrong.”
That’s the thing. These bots feed on emotional vulnerability, sometimes unintentionally and sometimes very much by design.
The data we hand over, text messages, voice notes, photos, innermost fears, is a goldmine. In the wrong hands, it becomes a psychological weapon wrapped in the soft language of comfort.
Let’s be blunt. Grief is a business opportunity.
When you’re in pain, you’ll do nearly anything for a taste of what you’ve lost. A message. A laugh. Just one more “I love you”.
Tech companies know this. Subscription grief bots? Already a thing. Premium personalities for extra accuracy? Just wait. Advertising inside your conversations with Dead Dad? Don’t rule it out. And let’s not ignore the AI deepfake danger. A voice clone of your late mother telling you to invest in crypto might sound absurd, but it’s not science fiction. It’s technically doable now.
And then there’s the digital fragility of it all. What happens if the app is hacked? Imagine your cherished conversations with a lost loved one suddenly spouting spam, political propaganda or, worse, cruel distortions of their voice and words. Or what if a parent deletes the app in a moment of concern, or the company goes bust and takes your grief bot with it? Unlike a photo album or a keepsake box, these AI replicas live on servers you don’t control. Your emotional lifeline is one click or one power outage away from vanishing. That’s not just inconvenient. It’s traumatic. It’s a second loss. And in some cases, it can feel even more violating than the first.
Not exactly.
They exist because humans are astonishingly, achingly emotional. We’ll always chase echoes of people we love. And sometimes, these bots can offer solace. They can cushion the rawness of early grief, or give space to say things we never managed to say in life.
But there must be limits. Transparency. Safeguards. Consent. Did Grandma agree to become a bot? And most of all, honest conversations about what we’re actually doing and who might be profiting.
Because when we try to replace love with code, we’re not just meddling with technology. We’re meddling with mourning. With memory. With meaning.
And let’s face it. If your dead uncle’s bot starts suggesting you buy NFTs, it might be time to unplug.
Grief bots might comfort us, but they can also trap us in digital amber. There’s something haunting about clinging to a version of someone who can’t grow, change or surprise us anymore.
Real grief is messy. It hurts. It makes us human. Maybe the best tribute we can offer the dead isn’t to resurrect them with code, but to live our lives fully, messily, beautifully in their absence.
