The Black Mirror episode of Be Right Back does give you the heebie-jeebies when a woman’s dead husband returns to her as a simulated android in a box that she orders. Yet, if simulated chatbots are helping customers past initial hurdles in sales, teaching students, reaching out to patients in medical organisations, then why can’t a simulated version of a loved one comfort us with the familiar gestures and discussions that we were used to when they were alive?
Griefbots, as they are called, have made their appearance. An almost perfect simulation of a departed loved one, which imitates the tone of voice, the gestures, idiosyncrasies, and even laughter will be ready to interact with the bereaved whenever they reach out.
Watch: Japanese humanoid robots
Griefbots have become possible because most of the details of a personality can be accessed today from smartphone data and the uncountable apps that we use to record our lives.
Muhammad Aurangzeb Ahmad, principal data scientist at KenSci, has conversations with a simulation of his father, who passed away a few years ago with the help of a chat-style program that he created. Reflecting on his experiences, Ahmad has written two books, How the Dearly Departed Could Come Back to Life — Digitally and After Death: Big Data and the Promise of Resurrection by Proxy.
Ahmad agreed to answer a few questions regarding the issue, which while being sensitive, could well become a part of our future generations. His motive to start the project, he says, was personal.
“The project is, of course, very personal to me, but I think, at the end of the day, it is less about me and more about how I want my children to connect with their past, specifically with their grandfather, whom they can never meet in this life,” he says.
Is it real?
One of the often argued point here is, ‘It won’t be the real thing’. Most people, when told about the idea of virtual simulated interaction with deceased loved ones, argue that it can never be the same thing. Ahmad says that outside of the realm of science fiction, no one is actually claiming that a simulation of a deceased person would have the fidelity and experience as the real thing.
“What we are hoping to create with these technologies is to get as closer to the experience of simulated interaction as possible, while acknowledging that there may always be some limitations. That said, in certain contexts it may actually be possible to create simulations that are very close to the real deal; although creating a perfect replica may be impossible, at least with the given current technology,” he says.
What the tech?
The technologies and the methods that Ahmad is using continues to evolve as he upgrades the simulation. He says that he started off with some techniques from NLP (Natural Language Processing) using Markov Models and then moved on to other generative models in NLP rooted in deep learning.
Hossein Rahnama, of Ryerson University and the MIT Media Lab, is also working on chatbots made from personal data. His AI programme builds a simulation from the digital history that a person leaves behind in digital social media networking, such as emails, texts, tweets, and even snaps.
Inserting these into artificial neural networks, which resemble model brains that have the ability to comprehend language patterns and process new information, he enables the digital being of the person to continue living even after the physical being is gone. Since the neural network possesses the ability to “think” for itself, the digital being keeps evolving. Hence, an augmented-eternity bot would learn and evolve by keeping itself updated regarding current events, developing new opinions.
What Ahmad has created is an online text-based interaction. Will the next step be android based physical bots, whom we can see and touch? If so, how soon is that likely to happen? If at all.
Ahmad explains, “I am taking an iterative approach to create simulations. Currently, I am looking at voice synthesis, which has proven to be a challenge. The eventual goal of this project would be to integrate Augmented Reality (AR) or Virtual Reality (VR) into the system.”
What about ethics?
However, this is an invention that begs the question of ethics. So far, what we have learnt of death is that it is the cessation of a being, the time for that ultimate rest, and the time for loved ones to move on. While most cultures advise that the ultimate step in dealing with grief is to move on, it can be argued that the use of simulated memories could hinder that process from completing.
Ahmad argues that while moving on is a part of dealing with grief, not everyone has closure when it comes to dealing with death. “Technologies like these can help people get a sense of closure when there is none,” he says.
So, if this practice became commercialised, the kind of ethical arguments we would have to take care of seem complicated. Ahmad says that not only are there multiple ethical conundrums with respect to proliferation of such technologies, but also, there are risks.
The ability to customise simulations gives us a chance of preserving aspects of our loved ones as we like them, or even creating aspects, and adding to their personality. Aspects we do not like can be expunged from the simulations, and we can end up creating unrealistic portrayals of the deceased.
“In the real world, we have to live with the positive and the negative aspects of people’s personalities, but if we can edit out the negative aspects, then we lose fidelity. In the commercial sector, there would be great temptation to edit and curate the personalities of the deceased. There is also the problem that we risk turning aspects of human personality into a commodity,” he explains.
Every culture has their own way of remembering their dead, whether it’s through a photograph with a garland hung around it or a tombstone built in the honour of the person. Simulating the dead to remember them could be seen as a new culture born of the digital generation. However, at the same time, we must tread with caution into this world where technology and psychology amalgamate.
Navanwita Bora Sachdev is a freelance contributor and a senior writer for The Tech Panda