Artificial intelligence to talk to the dead, an application full of questions

Carmen Rodriguez
Madrid, Jul 2 (EFE).- When a loved one dies we can remember them with photos, videos or messages and even ‘talk’ to them in our minds, but What if he was able to answer us? There is nothing supernatural about it, but rather another application of artificial intelligence (AI) not exempt from socio-ethical questions.

As if challenging death or promising something resembling eternal life, the digital afterlife industry offers the possibility of creating avatars of those who are no longer with us so that users can converse with them in real time, reproducing their image and voice, only the latter or in writing.

These ‘grief robots’ are created from the deceased’s digital footprint – content on social networks, messaging, emails – which is processed by a neural network that learns to imitate their behavior or way of thinking.

Each person has a way of experiencing grief and these technologies can change the way we deal with itBelén Jiménez, a doctor in psychology from the Universitat Oberta de Catalunya (Spain), told EFE that it is “essential to ensure respect and dignity for the deceased person as well as the psychological well-being of the user.”

In this area, he adds, there are no “all or nothing” answers, because it is a complex issue that depends on the technology, the grieving person and the designers of the tool, and there are still almost no scientific studies.

Conversational videos in which the person records answers to questions about their life, conversations with AI-generated chatbots with the voice of the deceased or only with text are what companies such as StoryFile, Eternos.Life, Decembre Proyect or You, Only Virtual offer to varying degrees.

In some cases, it is the person who prepares everything to leave this service active after their death, but it is also possible that it is their loved ones.

Digital afterlife industry

Jiménez, a specialist in how digital media mediate the experience of grief, points out that “we have prejudices regarding this type of tools and it is no wonder if we take into account that they are being created mainly by the growing digital after-death industry,” which seeks “commercial and economic objectives and not necessarily therapeutic ones.”

A recent study by the University of Cambridge analyses three hypothetical scenarios to show the possible consequences of a “careless design” of this technology without safety standards.

The research, led by Katarzyna Nowaczyk-Basińska, describes a scenario in which a grandmother’s bot makes an advertising recommendation to her granddaughter, another in which a child becomes distressed because the dead mother’s replica starts generating confusing responses.

In the third case, a father leaves an avatar for his children for 20 years. One begins to receive an avalanche of unwanted messages and the other ends up emotionally exhausted from interacting with him but is tormented by the fate of the chatbot.

Conversational videos in which the person records answers to questions about their life, conversations with AI-generated chatbots with the voice of the deceased or just with text are what companies such as StoryFile, Eternos.Life, Decembre Proyect or You, Only Virtual offer to varying degrees. EFE/Gema García

Dependency relationship

Bots can be designed to risk making the mourner believe that the answers they get are really from their loved one.which could create a relationship of dependency or even “the suffering caused by a ‘second loss’ if the device disappears, for example due to technical problems,” warns Jiménez.

‘Eternal You’, a documentary presented this year at the Sundance Film Festival (USA), followed the beginning of the grief bots and some of the first users, who “seek comfort with these applications and would give anything to be able to speak to their deceased loved ones one more time.”

This is explained on the festival’s website by its directors, Hans Block and Moritz Riesewieck, who also interviewed executives from these companies and addressed the current problem of dealing with grief.

“There were things that scared me and a lot of things that I didn’t want to hear, that I wasn’t ready to hear,” one of the users says in the trailer.

A bot, Jiménez says, can produce responses that are not consistent with the user’s memories of the deceased, which can cause frustration or pain.

The feeling of a second loss “when one realizes the lack of emotional authenticity” that characterized the relationship with the deceased or can somehow alter their memory, “which can lead to uneasiness.”

South Korean television channel MBC made a series of documentaries about immersive reality experiences with virtual recreations so that a mother could spend a few hours with her deceased seven-year-old daughter or a man could reunite with his dead wife, both of whom expressed their gratitude for the experience.

These ‘thanatotechnologies’ not only transform the experience of grief “but also the very concept of mortality or even ‘resurrection’,” says Jiménez, for whom traditional religions will coexist “with new and growing ways of understanding the world where digital immortality is an essential element.”

We are -he warns- facing a new technological development, based on AI, “of high risk that must be regulated”, since it involves very complicated socio-ethical challenges.

Each country has its own rules on data protection. In the case of the European Union, if such applications are directed at its citizens, the European AI Regulation and the General Data Protection Regulation would apply.