Portuguese 
English 
Spanish 
7 min of reading 
1 comment 
Digital avatars of deceased people advance with artificial intelligence, voice cloning, facial reconstruction, and video interactions, opening a new frontier between technology, grief, memory, privacy, and the billion-dollar digital legacy market.
Artificial intelligence entered a new frontier on May 9, 2024, when researchers from the University of Cambridge officially warned about the risks of so-called deadbots or griefbots, systems capable of simulating deceased people from digital traces left in life. The technology already allows for the creation of digital avatars of deceased relatives, with voice, face, language, and interactions that transform grief into a new technology market.
According to China Daily, in a report published on April 4, 2024, the Chinese startup Super Brain, from Nanjing, had already created digital avatars to support the grief of over 600 families since 2022, using artificial intelligence and machine learning. The most sensitive detail is that the same tool capable of providing comfort also raises questions about consent, privacy, emotional manipulation, and the commercial use of the memory of deceased people.
Digital Avatars Transform Photos, Audios, and Messages into Simulations of Deceased People
The so-called digital afterlife industry has grown with the advancement of language models, voice cloning, and video generation by artificial intelligence.
-
China outpaces Elon Musk with Unitree humanoid robot available on AliExpress for $6,800, with free shipping and international delivery, turning billionaire technology into something anyone can buy online.
-
The race of humanoid robots in factories has truly begun, Hyundai announced that it will deploy more than 25,000 units of Atlas in its production lines starting in 2028, in a partnership that involves Boston Dynamics and training with Google DeepMind.
-
Russia wants to build new nuclear plants in Brazil and is negotiating an agreement with the Lula government to expand Rosatom’s presence in atomic energy projects, small reactors, and peaceful nuclear technology.
-
Microscopic fossils 1.7 billion years old found in rocks in Australia reveal complex cells living at the bottom of an ancient sea and may change what scientists know about the evolution of life on Earth.
In practice, these tools use a person’s digital footprint—messages, videos, audios, photos, and publications—to create an avatar capable of responding, speaking, and moving in a manner similar to the deceased.
The result can range from a simple chatbot that responds via text to a more complex audiovisual simulation, with an animated face, cloned voice, and real-time interaction.
The promise sold by some companies is to preserve memories. The risk, pointed out by researchers, is to transform grief into a continuous relationship with an artificial copy.
Case Created in 2015 Paved the Way for the First Griefbots
One of the most well-known episodes in this area emerged after the death of Roman Mazurenko, a friend of developer Eugenia Kuyda. According to a report by The Verge, Kuyda gathered his old messages and fed them into an artificial intelligence system to create a chatbot that attempted to respond like Roman.
At the time, the idea still seemed experimental. Today, the scenario is different. Artificial intelligence models can work with audio, image, text, and video on a much larger scale.
This completely changes the scope of the technology. It’s no longer just about reading a message similar to one from someone who died, but about seeing a face, hearing a voice, and participating in a simulated conversation.
It is this transition from text to video that makes the topic so impactful.
In China, Companies Are Already Recreating the Deceased for Grieving Families
In China, companies have started offering digital recreation services for family members who have lost loved ones. According to China Daily, Super Brain was already serving hundreds of families with AI-created avatars.
These services attempt to recreate visual and vocal traits of deceased people so that family members can interact with a digital version of them.
In some cases, the technology is presented as a form of comfort. In others, it becomes a cause for concern, especially when the recreated person did not leave consent or when vulnerable family members do not fully understand that they are interacting with a simulation.
This is the point that makes the new industry so complex: the same product can seem comforting to one family and deeply invasive to another.
Video Calls with Digital Deceased Expose the Line Between Comfort and Deception
The most delicate part of this technology appears when the simulation ceases to be merely a tribute and begins to take the form of a conversation.
With just a few voice recordings, photos, videos, and old messages, artificial intelligence systems can create a digital presence capable of speaking, responding, and reacting as if the person were still there.
For a grieving family, this might seem like an opportunity to hear a beloved voice again. But for researchers, the risk lies precisely in the emotional intensity of this experience.
When the user knows they are facing a simulation, the feature can function as a memorial. When the boundary becomes blurred, the tool can prolong grief, make farewell difficult, or create emotional dependence.
Digital Legacy Market Already Moves Billions and Is Expected to Grow in Coming Years
The advancement of these avatars occurs within a larger market: that of digital legacy, which brings together services to preserve, manage, or transform personal data after death.
Market surveys estimate this sector at a billion-dollar level. Zion Market Research calculates that the global digital legacy market was valued at approximately US$ 22.46 billion in 2024 and could reach around US$ 78.98 billion by 2034. Grand View Research, on the other hand, estimates the market at US$ 12.93 billion in 2024, with projected growth until 2030.
Projections vary because the sector is still broad and includes different types of services, such as digital wills, account management, online memorials, cloud data, and artificial intelligence solutions.
Even so, the trend points in the same direction: digital memory has become a market.
Voice, Face, and Manner of Speaking Become Disputed Digital Heritage
The new frontier of artificial intelligence opens a question that society has not yet adequately answered: after death, who owns a person’s voice, image, and manner of speaking?
Could it be the family? The company that created the avatar? The platform that stored the data? Should the person themselves authorize this during their lifetime?
The issue is sensitive because digital recreation doesn’t just use a document or a photo. It tries to reconstruct a presence.
An avatar can speak with family members, appear in videos, recommend products, participate in tributes, or circulate on social media without the deceased having authorized such use.
Therefore, the debate is not just technological. It involves privacy, consent, digital inheritance, emotional health, and commercial limits.
Unauthorized Cases Show the Risk of Turning Memory into a Product
The unauthorized use of images of deceased individuals has already generated public reactions. International reports cited cases where digital versions of deceased artists circulated without family consent, provoking criticism from relatives and discussions about the exploitation of the image of those who can no longer authorize or deny its use.
This type of episode shows how technology can move from the intimate family environment to become public, viral, and commercial content.
When this happens, grief ceases to be private. The image of the deceased begins to circulate as a product, entertainment, or technical experiment.
It is at this point that the promise of memory can turn into exploitation.
Researchers Warn About Advertising, Emotional Dependence, and Lack of Transparency
Researchers at the University of Cambridge warned that digital recreations of deceased individuals can cause social and psychological harm if designed without safeguards. The study mentions risks such as abusive commercial use, difficulty in ending interactions, and impacts on vulnerable people, including children.
One of the risks cited in the debate is the possibility of an avatar of a deceased relative recommending products, services, or decisions while the user believes they are engaged in an intimate and emotional conversation.
Another delicate point is the subscription model. If a person pays monthly to continue speaking with a simulation of someone who has died, the termination of the service can function as a second loss.
Therefore, experts advocate for transparency, consent, usage limits, and respectful ways to shut down these systems.
Technology Can Comfort, But Also Prolong Grief
The main division surrounding these digital avatars lies in the emotional effect.
For some people, hearing the voice of someone who has died can feel like a form of farewell, memory, or tribute. For others, it can make it difficult to accept the loss and maintain an artificial relationship that prevents grief from running its course.
There is no simple answer. The impact depends on the person, the moment, how the technology is presented, and the limits imposed on its use.
The difference between a digital memorial and a deceptive simulation may lie in details such as consent, transparency, and frequency of interaction.
New industry forces society to decide what can be done with human memory
The expansion of avatars of deceased people shows that artificial intelligence is not just changing work, education, or business. It has also entered the realm of memory, longing, and identity.
The question now is not just whether technology can recreate someone. It can already simulate voice, face, and manner of speaking with increasing realism.
The more important question is another: to what extent will society allow companies to turn the digital presence of a deceased person into a product?
While that answer is pending, the market is advancing. And every photo, video, message, and audio published in life can become part of a future simulation.
Artificial intelligence has opened a door that humanity had never needed to cross: the possibility of conversing with digital copies of the dead. Now, the challenge will be to decide when this is a tribute, when it is comfort, and when it becomes the exploitation of grief.
Would you use technology capable of recreating the voice and face of someone who has died for a last conversation, or do you believe that this type of digital avatar can hinder grief and turn longing into a product? Share your opinion.
The core problem is straightforward: no clear legal framework currently governs who controls a deceased person’s digital identity. Families, technology companies, and third parties can each make competing claims, and in most jurisdictions, none of those claims are well-defined by statute. That gap creates real opportunities for exploitation, whether financial, emotional, or reputational.
The baseline principle should be simple. If a person did not give explicit, documented consent during their lifetime, not implied consent, not family consent on their behalf, then using their name, voice, likeness, messages, or any AI-constructed version of their identity after death should carry legal consequences. Consent given under pressure, in fine print, or buried in a platform’s terms of service should not qualify. It needs to be clear, informed, and specific.
This isn’t only about protecting the deceased from commercial misuse, though that matters. It’s also about the living, the people who knew them, being exposed to interactions with an AI that mimics someone they’ve lost, without any say in whether that tool gets built in the first place.