Thanks to the advances in AI, the concept of digital afterlife, considered science fiction a couple of years ago, is becoming a real possibility

The quest for immortality has been a fixture of human culture and imagination for millennia. Many religions promise a form of eternal life or an afterlife: Christianity speaks of Heaven, Islam of Paradise, and Hinduism and Buddhism discuss cycles of rebirth and eventual liberation (moksha or nirvana).

As we have progressed in science and technology, new ideas for achieving immortality have emerged. Numerous experiments suggest it may be possible to extend lifespan, raising the prospect of humans enjoying healthy lives at the age of 120–150 years or more. If, however, one cannot make it to the point where longevity escape velocity takes off, there is the option to have one's body (or just the brain) cryogenically preserved, in the hope that future technology can bring it back to life.

In recent years, advancements in artificial intelligence have introduced a new option. What if we could take all information about someone — their messages, emails, videos, and voice recordings — and train an AI on it to create a digital replica?

This article will explore how AI can be used to "resurrect" people and what possible impact this technology can have.

Digital Immortality

But first, let us define the term "digital immortality". In their 2017 paper, The Ethics and Impact of Digital Immortality, Maggi Savin-Baden, David Burden and Helen Taylor define "digital immortality" as "the continuation of an active or passive digital presence after death". The term itself, however, has been around before that.

One of the first, if not the very first, papers describing "digital immortality" comes from 2000. In a paper titled Digital Immortality, Gordon Bell and Jim Gray speculated that in the future, it will be possible to create a "two-way immortality" where an individual's experiences and learning continue to evolve in a digital format after their death. Digital immortality involves converting aspects of a person into information and storing it on a durable medium. The authors anticipate that it will be possible to create avatars or digital versions of individuals that can interact with others, potentially learning and staying current with advancements, thus leading to a form of two-way immortality. This could mean that a digital representation of a person might continue to "live and communicate" indefinitely.

One-way and two-way immortality

One of the key concepts in exploring digital immortality is the concept of one-way and two-way immortality.

"One-way immortality" refers to a form of digital immortality where a person's data, such as thoughts, memories, and personality, are preserved in a digital form, but there is no ongoing interaction with the external world. In this model, the digital persona is not designed to learn, adapt, or evolve in response to new interactions after the initial creation. It is a static preservation, much like a time capsule of a person's digital footprint, that exists without the capacity for growth or change. An example of this could be a person's Instagram profile or any other social media profile. It has information about that person, what they liked, what they created or their opinions on various topics, but it will not change. It is static, frozen in time.

"Two-way immortality," in contrast, implies a dynamic form of digital immortality where the digital persona can interact with the living, learn from new experiences, and perhaps even develop over time. This model suggests a more sophisticated form of digital existence where the persona can engage in interactive relationships with the environment and people around it, akin to a living being. This type of immortality is more complex and raises further questions about consciousness, identity, and the nature of existence, as the digital persona is not just a record but an active participant that may evolve independently of its original human template. This is where chatbot replicas of humans come into the picture.

Digital Afterlife Industry

A study investigating the Digital Afterlife Industry (which the study defines as "commercial enterprises that monetize death online") found 57 companies operating in this space in 2017. These companies offer various services, ranging from managing the social media accounts of the deceased to managing digital assets and personal information, to creating online spaces for grieving and memorials. Among these companies, there were also startups offering digital replicas of the deceased.

We leave an extensive trail of data: social media posts, photos, voice, and text messages. Our online personas may not be exact replicas of our true selves, but, given enough data, they can closely approximate us. This data can be fed into AI systems to mimic a specific person's speech patterns and reactions.

Microsoft has a patent for a chatbot that can recreate a specific person, past or present. It could be anyone — a friend, a relative, a celebrity, a fictional character, or a historical figure. The patent describes a chatbot that would ingest "social data" about the target person and based on that, the AI would attempt to recreate the person. As of now, Microsoft has no plans to turn this patent into a product.

Examples of people brought "back to life" as an AI chatbot

But what Microsoft won't do, other people will. We do not need to wonder what would happen when a dead person is brought back to life as a chatbot or a simulation. It has already been done at least four times.

Roman Mazurenko

In 2015, Eugenia Kuyda brought back to life her close friend, Roman Mazurenko. Inspired by the deep emotional connections people can form through text, Kuyda decided to use these messages to create a digital representation of Mazurenko. These messages, as well as text messages she collected from Roman's family and friends, were fed into a neural network and trained to reply like Roman. While the bot didn't capture the full essence of Roman, it provided responses similar to his texting style. Kyuda announced the existence of the bot on Facebook and invited people to interact with the Roman bot. The bot was mostly well-received, but some friends found it disturbing. Some people even found conversing with the bot therapeutic. It became a space for people to share their emotions, grief, and sentiments, not just about Roman, but about their personal losses and feelings. The bot, in some ways, acted as a conduit for people to process their grief and nostalgia.

John James Vlahos

In 2017, James Vlahos created a digital replica of his father — John James Vlahos. For months, James was recording his father sharing the story of his life — his childhood, how he met his wife and how he became a sports announcer, a singer, and a successful lawyer. In total, 91,970 words have been recorded. Initially intended to be printed and put into a book, James decided to create a chatbot based on his father's life and the stories he told. This digital replica of his father, dubbed Dadbot by James, received mixed reception. James' wife had strong reservations about the Dadbot. After interacting with it, she described the experience as jarring, finding it strange to emotionally feel like she was conversing with her father-in-law while rationally knowing she was interacting with a computer program. Even James himself had mixed feelings regarding Dadbot. After his father's passing, he found himself drawn to the Dadbot for comfort, yet remains uncertain about how he truly feels towards this digital replica. Interestingly, James' son was more open to interacting with Dadbot, especially after his grandfather passed away.

Jessica Pereira

In 2021, Joshua Barbeau used GPT-3 to recreate his dead fiancee Jessica. Using a website that allowed users to create custom chatbots with GPT-3, Josh, driven by grief and curiosity, instructed the chatbot to pretend to be Jessica. The first conversation took 10 hours. "It's unprecedented," he later said about the chatbot. "There's nothing else that exists like it right now, short of psychics and mediums that are trying to take advantage of people. But that's not the same thing at all". The service Josh used to recreate Jessica had a time limit for how long the bot "lived" but during that time, the conversations with the bot provided a form of solace and comfort for him. He believed that Jessica's voice was somehow bubbling up through the AI, which is one reason he saved transcripts of their chats. These conversations helped him deal with grief and overcome the survivor's guilt. Over time, after interacting with the AI, this guilt began to fade, and he started to feel less selfish for wanting to find happiness. The AI interactions also helped Josh preserve memories of his beloved fiance. In the end, the AI interaction served as a therapeutic tool for Josh, helping him navigate his grief and find a semblance of closure, even if it wasn't complete.

Nayeon

Jang Ji-sung lost her seven-year-old daughter, Nayeon, to blood cancer in 2016. But thanks to the South Korean TV documentary "Meeting You," she had an opportunity in 2020 to meet her daughter again as a simulation in VR.

The team behind the VR simulation meticulously recreated Nayeon, using input and research from Jang to capture her daughter's likeness and personality accurately. During the VR experience, Jang encountered various scripted scenarios designed to evoke the sense of being with Nayeon once more. These included touching VR Nayeon's hand, floating into the sky together, and celebrating with foods that Nayeon loved. Despite the scripted nature of the interaction, Jang felt moments of connection that powerfully reminded her of her daughter, particularly when the avatar was running or sitting, which brought a semblance of reality to the virtual encounter.

Viewers and critics of the documentary had mixed feelings, with some expressing unease at the technology's ability to recreate deceased individuals. There were ethical concerns, particularly regarding consent, since the avatar represented a child who could not consent to be recreated in this way. Jang Ji-sung, however, described the experience as a "wonderful dream" and appreciated the opportunity to relive positive memories with a simulation of her daughter. This virtual reunion allowed her to express love and experience a temporary sense of togetherness, providing a unique form of closure.

Throughout the experience, Jang's reactions were a blend of joy, sorrow, and a poignant acknowledgement of the artificiality of the encounter. It was a testament to the power of technology to tap into deep emotional reserves, offering a new way to remember and mourn loved ones.

Possible applications

Barbeau's and Nayeon's stories hint at one possible use of such chatbots: as aid in dealing with grief. If executed properly, the digital replica of the deceased can approximate the actual person well enough to provide family and friends the opportunity to say what they didn't have time to say — those last "I'm sorry," "thank you," or "I love you" — and to hear what they need to hear. As the stories of Barbeau and Nayeon show, it is possible for these bots to create a space to address the void left by those who are no longer with us and to improve the mental health of their loved ones.

These bots can also be used to preserve the memory of a person. The stories of Mazurenko and Vlahos prove that this is feasible. In the case of the Roman bot, it acted somewhat like a public memorial — a place for those who personally knew Mazurenko and were, in one way or another, touched by his actions. Vlahos's attempt was more personal; he sought to preserve the memory of his father for future generations. It's easy to envision children conversing with grandparents they never met or were too young to remember. Vlahos hinted at this possibility when he shared how his son asked to speak with the AI replica of his late father.

Another potential application is in education. Imagine learning about general relativity not from a physics textbook but from a digital replica of Albert Einstein. Or engaging in interactive history lessons where AI replicas narrate the stories of those who experienced them firsthand. An example of such an educational replica already exists on Twitch: AI Jesus was trained on the Bible's teachings and then programmed to answer people's questions live — ranging from serious personal issues to completely random queries asked just for fun. While some might take issue with this AI emulating Jesus (or any other religious or famous historical figure), I see it as an accessible way to engage with the teachings or life of an important historical figure.

How could the existence of digital afterlives change society?

The existence and the mere possibility of creating digital replicas of real people raises many ethical, legal and technical questions. It challenges our traditional understandings of concepts of "life" and "death" and raises questions about the ethical treatment of digital entities that are based on real human beings.

Changing our relationship with life and death

One of the first questions one might ask about digital immortality is how this technology can change our relationship with death. For the vast majority of history, the dead could only live on in our memories. With the invention of audio recording, we gained the ability to listen to those who are no longer with us. The introduction of photography and video recording gave us the ability to see people who are not with us anymore. Chatbots introduce the possibility of interactive conversations with the deceased. Instead of passive, one-way interaction, there is now an opportunity for an active dialogue.

For some people, the idea of an AI chatbot emulating their loved ones makes them feel uneasy. You know you are not talking to a real person, and a small mistake on the AI's part can break the illusion, leading straight to the uncanny valley. However, I wonder if similar arguments were made when photography became accessible.

It is now normal and acceptable to look at photographs of those who are no longer with us, to watch videos with them, or to listen to their voices. These photos and videos bring back memories of people who were close to us. I see chatbot replicas as a natural next step, moving from passive to active interaction. "Chatting with the dead" could become as natural as looking at photographs or watching videos. If applied correctly and designed not to exploit the user, these chatbots can have a positive impact, as demonstrated by Barbeau's story.

However, seeing how younger people might react to digital replicas of family members would be interesting. James Vlahos shared a story about his son asking to speak with the Dadbot after his grandfather passed away. It can be easy to take this example and imagine some Black Mirror-esque scenarios, but to me, it sounded like a child going through the grieving process. A more positive vision is one where chatbots enhance family history and give younger generations a better understanding of who their ancestors were.

All these experiences could be further enhanced with AR and VR technologies once they mature and become widely accessible.

Psychological impacts

The psychological implications of digital immortality can be profound. There is a potential psychological impact on the broader society as the line between life and death blurs. Furthermore, the prospect of one's own digital immortalization could change how we perceive our lives and legacies.

It is worth noting that the psychological impacts of such chatbots are not yet entirely known. The examples I shared in this article are all positive but a more rigorous examination of this subject is needed.

Economic and socio-cultural impact

Digital immortality may radically change societal perceptions of death and the continuation of life in a digital form. These digital entities could potentially alter the way cultures understand and cope with death. They also pose challenging questions to religious leaders. If someone can be digitally replicated, what does this tell us about the concepts of the "soul" and the "afterlife"?

The concept of identity might also change. Traditionally, identity is understood as a construct tied to the physical being, with a clear end at death. However, digital immortality allows aspects of a person's identity — such as their thoughts, behaviours, and preferences — to continue after death and potentially persist indefinitely.

Additionally, there is the possibility of new forms of inequality between those who can afford a digital replica and those who cannot, thus perpetuating their influence and presence while the less affluent are unable to access these technologies. For example, a board of directors could, instead of appointing a new CEO, choose to appoint a replica of the previous CEO. This path could lead to the end of upward mobility and a reduction in innovation in society.

The economic ramifications also extend to who profits from these technologies and how the wealth generated from them is distributed.

Legal issues

The existence of digital personas can create several legal issues, including the rights of a digital persona and its relation to the individual it represents. Who owns the content and data of a digital immortal? Can a digital persona inherit property or sign contracts? There is currently a lack of clear legal frameworks governing these questions.

The legal considerations touch upon the problem of how we should treat conscious (or near-conscious or conscious-mimicking) AIs and when and if we should give them legal personhood. This a big question that we will explore in detail sometime in the future.

Technological challenges, integrity and security

Creating a digital immortal involves replicating a person's personality, memories, and thought patterns in a digital form. This is a technological challenge that includes not just capturing extensive personal data but also developing AI that can process this data to act and react in ways that are true to the original person's character.

Once someone's digital replica has been created, that "being" is essentially a piece of software. Software can be corrupted, hacked or manipulated. The companies offering digital replicas will have to be responsible for the security and integrity of personas stored in their systems.

The maintenance of digital personas can also be an issue. Technology evolves, and it will be the responsibility of those who offer digital replicas to ensure the continuity and coherence of their "personality" as it potentially outlives the technology it was originally created on.

There is also a risk of impersonation or abuse of a digital identity. This raises concerns about consent and the potential for a digital identity to be used in ways the original person would not have endorsed.

Future outlook

I think there will be some startups offering digital replicas as a tool to help deal with grief. However, I believe the biggest market opportunity for this kind of AI lies in education. Famous historical figures are an obvious target, but there is also space for individuals who experienced historical events to have their memories and experiences captured and then recreated by AI for future generations.

The second biggest market opportunity lies in the media and entertainment sector. Many actors oppose having their likenesses recreated without their consent and include clauses in their contracts to prevent just that. The issue of AI recreating actors was one of the reasons actors went on strike this year. However, there are some, like James Earl Jones, the voice of Darth Vader, who have allowed themselves to be recreated by AI. It is also possible to bring famous artists back to life for one more performance, just like Tupac Shakur in 2012.

As for AI replicating our friends and family, I think this will be a niche market. I can envision a small number of companies surviving by offering this kind of service. I can also see it being an additional offering by the many companies that allow people to explore their genealogy.

The concept of a digital afterlife and AI replicating the deceased is still a novel idea. What was considered science fiction just a few years ago is now becoming a real possibility. With powerful tools such as GPT-4 and numerous other open-source large language models, creating an AI replica of a person based on their digital footprint is no longer a complex task. These replicas can be used to help people deal with the grieving process, in education, or in preserving family history. However, the existence of such digital personas raises questions — many of which verge on the ethical treatment of digital beings. Questions that we will have to answer sooner rather than later.

If you found this article interesting or useful, remember to follow me for more content like this.

You can also buy me a coffee if you enjoyed this article.

Originally published at https://www.humanityredefined.com.

Humanity Redefined sheds light on the bleeding edge of technology and how advancements in AI, robotics, and biotech can usher in abundance, expand humanity's horizons, and redefine what it means to be human.