AI memorials are changing the way people remember loved ones who have passed away. Imagine asking your late grandfather what his childhood was like and hearing an answer in his familiar voice. For families who have lost someone, being able to revisit personal stories may offer comfort. But hearing someone’s voice again does not mean having that person back.
When a Familiar Voice Starts Making Things Up
Not all AI memorials work in the same way. HereAfter AI, for example, allows people to record their life stories while they are alive. Family members can later ask questions and hear the person’s original recordings in response. Other AI tools go further by generating new conversations based on a deceased person’s voice, messages and personal information.
This creates a different problem. An AI replica may sound like someone’s loved one, but its answers are not necessarily things that person ever said or believed. A family member asking for advice may receive a convincing response generated by AI rather than a genuine memory.
A person’s voice can be preserved, and their stories can be revisited. But an AI-generated answer should never be mistaken for the person’s own words.
The Person Who Died Should Still Have a Say
There is another question families may not consider before creating a digital replica: did the person want to be brought back in this way?
Having someone’s photographs, voice messages or private conversations does not automatically mean their family has permission to recreate their identity. Some may welcome an interactive record for future generations, while others may prefer to keep their personal lives private.
This is why consent should ideally happen before death. People should be able to decide whether their likeness can be recreated, who can interact with it and whether the digital version can generate new responses. If they never agreed to it, families and companies should not assume that access to their data means permission to reproduce them.
When Remembering Someone Comes with a Monthly Bill
The emotional side becomes even more complicated when digital remembrance turns into a subscription service.
Companies need revenue to maintain their platforms, store recordings and operate AI systems. Charging for these services is not automatically exploitation. However, there is a difference between paying for data storage and paying repeatedly to maintain access to a digital version of someone you have lost.
If cancelling a subscription means losing access to a familiar voice, some families may struggle to stop paying. This is where the idea of grief as a subscription becomes uncomfortable. Businesses should not design emotional experiences that make users feel guilty about leaving or encourage them to keep returning by suggesting that the deceased still misses them.
Let Memories Stay, Not the Illusion
AI memorials do not need to disappear. They can help preserve family histories, save voices that might otherwise be forgotten and give future generations access to stories they would never have heard.
But the technology needs boundaries. People should have control over how their identities are used, AI-generated responses must be clearly labelled, and families should be able to download or delete their data without being trapped by ongoing subscriptions.
Perhaps the most meaningful thing we can leave behind is not an AI version of ourselves that keeps talking, but the stories, memories and relationships we shared while we were alive.
AI can help families remember someone. It should not decide who that person was, speak on their behalf without permission, or turn the difficulty of letting go into a business opportunity.



