Technology can now do something once reserved for memories and imagination: it can bring a person back to life on a screen.
With just a few clicks, AI can recreate someone’s voice, face, expressions, even the way they move. For some, it feels like a miracle. Families can “see” a loved one again, hear them speak, relive a moment they thought was gone forever. It can be comforting in a way that old photos and shaky phone videos never quite are.
However, there’s a darker side we don’t talk about enough. Who gets to decide whether a person is digitally resurrected? Is it a tribute, or is it exploitation when the one being recreated can’t consent? The line between honouring someone and using their image for entertainment (or profit) is becoming increasingly blurry. As AI becomes more powerful, the real question isn’t can we bring people back on screen. It’s should we?
It could help grieving families process loss.
Some people find comfort in AI recreations that let them hear a loved one’s voice again or see them one more time. For those struggling with sudden loss, it might provide closure that traditional grief counselling can’t offer.
That’s why completely banning the technology seems harsh to families who find genuine healing in it. However, mental health professionals should guide its use to prevent unhealthy attachment to digital recreations instead of processing actual grief.
Dead people can’t consent to how they’re portrayed.
Creating videos of someone who’s died means putting words in their mouth without their permission. They can’t object to being made to say things they’d have hated or appearing in contexts they’d have refused. It has to be firmly established that estates and families have legal authority over deceased people’s digital likenesses. This protects against exploitation whilst allowing controlled use by those who actually knew them.
It opens the door to massive misinformation.
AI videos of dead politicians, celebrities, or historical figures could spread false information that reshapes public memory. People might genuinely believe fabricated speeches or statements because the technology looks so convincing. This becomes especially dangerous with recent deaths where people’s memories are still forming. Mandatory labelling of AI-generated content featuring deceased people could help prevent historical revisionism.
Families might disagree about whether it’s appropriate.
One child might want an AI recreation of their dead parent for comfort, while siblings find it deeply disturbing. These conflicts could tear families apart over whether keeping digital versions violates the deceased’s dignity. Because of this, advance directives about digital legacy should become standard. People need to state whilst alive whether they consent to being recreated digitally after death, removing the burden from grieving relatives.
Commercial exploitation becomes inevitable.
Companies will absolutely use AI to put dead celebrities in adverts, films, and products they never endorsed. The profit motive means someone’s image gets used to sell things, regardless of what they’d have wanted. We need legislation that treats digital likenesses as inheritable property for several decades after death. Estates should control and profit from commercial use, with unauthorised use facing serious penalties.
It could preserve important historical figures accurately.
AI recreations might help educate future generations by bringing historical figures to life in ways textbooks can’t. Hearing Churchill or Einstein explain their actual work in their own voice could make history more engaging and accessible. Of course, this only works if the content is historically accurate and clearly labelled as recreation. Educational use should require rigorous fact-checking and expert oversight to prevent propaganda disguised as history lessons.
The technology doesn’t understand dignity or respect.
AI has no concept of what would embarrass, upset, or betray a deceased person’s values. It will generate whatever it’s asked to create without moral consideration of whether the person would have consented. Human oversight and ethical review boards are a must for this very reason. The technology needs guardrails built by people who understand context, dignity, and the weight of representing someone who can’t object.
Different cultures view death completely differently.
Some cultures believe disturbing the dead in any form brings bad luck or disrespects ancestors. Creating digital versions might be deeply offensive to communities with specific beliefs about honouring the deceased. Regulations would need to account for cultural and religious sensitivity. What feels healing to one family might be sacrilegious to another, requiring flexible approaches rather than one-size-fits-all rules.
Children might grow up with false memories.
Kids shown AI videos of dead parents or grandparents could form memories based on fabrications rather than reality. They might remember conversations that never happened or develop relationships with digital ghosts instead of processing actual loss. This could fundamentally alter child development and grief processing. Age restrictions and psychological guidance should govern children’s exposure to AI recreations of deceased family members.
It raises questions about identity after death.
If an AI version of someone exists saying and doing new things, at what point does it stop being them and become something else entirely. The philosophical implications challenge our understanding of personhood and mortality. Clear distinctions between preservation and creation matter legally. Recording actual footage differs fundamentally from generating new content the person never created, requiring different ethical frameworks.
People might refuse to let loved ones go.
Access to AI recreations could prevent healthy grieving by offering an illusion that the person hasn’t really gone. Some might choose digital relationships with the dead over forming new connections with the living. It could help if mental health professionals develop guidelines for healthy vs unhealthy use, but that’s murky territory. Occasional viewing for comfort differs from daily conversations with AI versions that prevent moving forward.
Public figures have different considerations than private citizens.
Politicians, celebrities, and historical figures already exist in public consciousness in ways private citizens don’t. The ethical bar for recreating them might reasonably differ from recreating someone’s grandmother. Public interest arguments might justify some uses that wouldn’t apply to ordinary people. However, even public figures deserve protection from being made to endorse positions they opposed whilst alive.
The technology will exist regardless of laws.
Banning AI recreations won’t stop people creating them, it’ll just push the practice underground where there’s zero oversight or ethical consideration. Outlawing something rarely works with technology that’s already widely accessible. That’s why regulation rather than outright bans makes more practical sense. Clear rules about labelling, consent, and permitted uses give society tools to address abuses without driving everything into unregulated spaces.
We need the conversation now before it’s everywhere.
This technology is advancing faster than our ethical frameworks can keep up. Waiting until AI recreations of dead people are commonplace means we’ve already lost the chance to shape how it develops. Society needs to engage with these questions immediately whilst we can still influence the technology’s trajectory. The choices we make now about consent, dignity, and appropriate use will define how we relate to death and memory for generations.
