Life after death is a topic that has kept some of the greatest minds occupied for millennia. It’s something we tend to shy away from talking about, but unfortunately, death and grieving are inevitable realities of being human. With ground-breaking research happening in Natural Language Processing and Artificial Intelligence, could griefbots hold the answer we’ve been looking for?
When you suffer loss, there is a longing to speak to the deceased again – and technology is making this possible. Death today comes with the uncanny nature of leaving behind a digital footprint, a legacy of social media posts, videos, pictures, and text messages – what are the living meant to do with these? Feed them into an artificial neural network and create a chatbot, or ‘greifbot’ version of the deceased, obviously.
These controversial projects are looking to how we can leverage deep learning to extend our lives by imitating an extension of life through a chatbot, or ‘griefbot’ – yes, you’re right this is very Black Mirror.
We’ve come a long way since ELIZA, Joseph Weizenbaum’s Computer program from the mid-1960’s that responded to questions like a psychotherapist by using predetermined phrases, simply repeating back the user’s questions in different formats to form an answer.
Advancements in artificial intelligence and natural language processing are allowing for more realistic and humanlike chatbot experiences. With a combination of greater computing power and enhanced deep learning algorithms, researchers have extended the layers of abstraction that can be processed by artificial neural networks. These networks are capable of allowing the software to identify and understand and process patterns in data such as image, sound, and text. For users of a griefbot, not only does this allow for greater understanding of their commands into the application, but also the output through the ability to imitate someone’s personality through their digital legacy.
Earlier this year, Artificial Intelligence start-up Luka’s co-founder Eugenia Kuyda garnered much media attention as her project to bring a deceased friend back through chatbot technology came to fruition. Kuyda had been working on a chatbot function that worked to recommend restaurants at the time. After Roman’s death, she decided to use texts between them, and some donated by friends and family to construct a Roman chatbot for everyone to interact with. Most of the users couldn’t believe how like Roman this chatbot was in phrases, humour, and even personality.
She is now focusing on Replika, a chatbot program that will learn through user interaction exactly how to communicate like the user. The ultimate personal assistant which can help with daily tasks, but one day carry on where you left off. You can sign up to have a griefbot of your very own.
Whilst Replika learns from direct and personal interaction, there is an issue with using data from more public communication, like social media accounts. We present ourselves online through a filter, creating an avatar of who we want to be and not who we really are. This means that social data used in creating a posthumous chatbot would not reflect a true you. This can be said for text messages too, think about what you’re like with different people in your life; work and home, partner and parent and your different interactions and digital legacy may reveal personal traits you wouldn’t want to expose to everyone.
We may just be at the beginning of our memorial-bot journey, but with demand will come supply and we’re bound to see more start-ups focusing on similar products.