HereAfter, an AI-powered chatbot, is a “posthumous messaging service” that aims to reconnect a deceased person with his/her living loved ones. Through preserving the “life story” of its users — enabled via an algorithm fed with details about their childhood, family, romantic relationship, career, and such — the chatbot can respond to questions in a simulated voice of the deceased user.
Microsoft and Amazon are among some of the tech giants treading into this new territory. In 2017, Microsoft filed a patent that would allow the company to reincarnate people as chatbots based on “images, voice data, social media posts, electronic messages” and other personal information. Apart from the typical mode of feeding the algorithm details of a person, the patent also states “the specific person may also correspond to oneself” (e.g. the user creating/training the chatbot) meaning a living person can train their own digital replacement ahead of their death.
Although no official plans and details were announced by the company since the filing, backlashes on social media ensued when the news of the patent’s approval broke out in December 2020, in which Tim O'Brien, General Manager of AI Programs at Microsoft, confirmed on twitter that “there’s no plans for this” — he also described the patented tech as “disturbing”.
Meanwhile, during the Amazon re:MARS conference this June, the company unveiled a new feature for Alexa, its cloud-based virtual assistant, that can speak in a deceased person’s voice, based on a short recording of that person.
Despite its noble intentions — that "while AI can't eliminate that pain of loss, it can definitely make their memories last", as described by Rohit Prasad, senior vice president and head scientist for Alexa — the new feature nonetheless met much of the same reception as Microsoft’s patented chatbot. Reactions toward the announcement are mostly negative or dismissive, such as “morbid”, “creepy with deep fake ghosts”, or “Day 2: Everyone gets dead grandma to swear”. With its cutting-edge technology under scrutiny, Amazon has yet to reveal the public launch date for Alexa’s new feature.
Amongst numerous iterations of the metaverse, Somnium Space, founded by Artur Sychov, sets itself apart with the forthcoming Live Forever mode. Inspired by the passing of Sychov’s father in 2017, the function will record the user’s movements and conversations through the Teslasuit — a full-body haptic suit for VR (which, by the way, has no connection to Elon Musk) — and replicate a digital avatar of a user based on the collected data, enabling people to talk with their deceased loved ones even when they have passed away.
“Literally, if I die — and I have this data collected — people can come or my kids, they can come in, and they can have a conversation with my avatar, with my movements, with my voice,” said Sychov. “You will meet the person. And you would maybe for the first 10 minutes while talking to that person, you would not know that it's actually AI. That’s the goal.”
To Sychov, the Live Forever mode is not only backed by an ideal vision of connecting the living and the dead, but also by the immense potential of VR in terms of data collection, where it can identify a person “more precisely than fingerprints”. Indeed, according to a study1 conducted in 2020, a VR system is capable of identifying 95% of the users from a pool of 511 participants, when trained in less than five minutes with tracking data of each participant. With the project currently in the works, the company aims to roll out the first versions of digital avatars, capable of basic movement and conversational abilities, for its users by next year.
Perhaps a more well-known example in the digital afterlife industry, in February 2020 South Korea’s Munhwa Broadcasting Corp released a documentary film titled Meeting You that depicts a VR experience for Jang Ji-sung, who lost her daughter Nayeon due to leukaemia four years ago. The experience started with Nayeon appearing from behind a pile of woods, galloping towards Jang. Tears streamed down Jang’s face as she reunited with Nayeon, where she reached out to stroke her daughter’s hair and embrace her — yet, in reality, she was wearing a VR headset and touch-sensitive gloves in a studio filled with green screens.
Whether holograms, chatbots, and VR doppelgangers are, or will be, effective ways for coping with grief is an inane question — since the expression of one’s sentiments could take on different forms, which does not necessarily hinge upon a “perfect” simulation of a dead person — it would be equivalent to challenging numerous kinds of traditional rituals and practices for the dead, such as the Jazz Funeral from New Orleans, the Jewish Sitting Shiva, or Japan’s Obon festival.
While information management and online memorial services such as Memories are relatively straightforward in terms of execution, digital afterlife services that involve technologies such as machine-learning and artificial neural network introduce new kinds of legal and ethical conundrums.
The first challenge pertains to the handling of data previously owned by a deceased person. Although a person can dictate in a will on how they wish to dispose or settle their digital assets, such personal wishes can be overridden by the executors or relatives of the deceased person in certain jurisdictions, say, the UK. A notable example is the will of renowned writer Franz Kafka2, where his lawyer and friend Max Brod ignored his fiduciary duties to destroy everything the writer left behind — letters, diaries, sketches, and more — and instead published some of the manuscripts by Kafka.
According to the UK’s Medical Research Council, the consent for decisions related to posthumous arrangements should be given out by the relatives with respect to the deceased person. Yet, in the case of digital afterlife technologies, does recreating a deceased family member — through using the digital assets they once owned — count as a disrespect towards them? Which party should the data protection law cater for, the living or the dead?
Such ambiguity also looms over the issue of human dignity when it comes to emulating a deceased person through chatbots and digital avatars. A more reliable approach could begin with the premise that digital remains of a dead person serve as a part of an “informational human body”3, meaning it constitutes to one’s personhood, which enjoys the right to be treated with respect worthy of a living human; or drawing from the Code of Professional Ethics by the International Council of Museums4, to treat the digital remains of a deceased person as an entity with inherent value, including their inviolable “human dignity”, instead of a means to an end.
However, putting aside commercial exploitation or anomalous intentions: how can one determine whether a deceased person’s dignity is intact, when the very existence of digital afterlife technologies entails treating their digital remains as a means to an end (i.e. to offer companionship or closure)? Furthermore, with the rise of superintelligent machines, how can one ensure emulations of the dead will align with our principles of human dignity?
In face of the multifold issues that comes with the digital afterlife industry, some questions will remain timeless and universal — ones that offer a reality check on what we cherish in a relationship and what makes our loved ones unique.
1. Miller, M.R., Herrera, F., Jun, H. et al. (2020). Personal identifiability of user tracking data during observation of 360-degree VR video. Sci Rep 10, 17404.
2. Morris, Arval. Law, Language, and Ethics. (1973). 73 COLUM. L. REV. 1342, 1344 (reviewing WILLIAM R. BISHIN & CHRISTOPHER D. STONE, LAW, LANGUAGE,AND ETHICS (Foundation Press 1972)).
3. Stokes, P. (2015). Deletion as second death: The moral status of digital remains. Ethics and Information Technology, 17 (4), 1–12.
4. International Council of Museums. (1986). Code of Professional Ethic.